Forgetting

• Forgetting is conceptualised as a failure to internalise information - 'forgetting is the inability to recall something now that could be recalled on an earlier occasion' (Tulving, 1974).
• In order to forget a memory, a memory must be formed beforehand.
• A failure in internalisation could be attributed to factors such as a lack of attention.
• There are two definitions of forgetting - unavailability and inaccessibility.

• Biologically, we might view memory as a change in connectivity between neurons - i.e. the connections degrade resulting in the memory being lost and unavailable.
• No retrieval if the connection ceases to exist. 
• Bailey and Chen (1989) - examined unavailability using sea slugs. New neural connections formed following learning. These connections degraded over time. Extent of degradation correlated with forgetting.
• This is incredibly hard to demonstrate in humans - scientists would 1. Have to see all neural activity/connectivity pre-learning, 2. All neural connectivity post-learning (i.e. the change in isolation of all other changes/learning), and 3. Degredation of that specific connection post-learning. In addition to this, the subject wouldhave to be a in a cognitive vacuum in order for the testing to work.
• As a result, it has been shown that unavailability cannot be demonstrated universally.

• Sometimes a cue is neeed in order to retrieve that memory - in the absence of that cue we have an issue of inaccessibility.
• Memory still exists biologically but needs a prime to access it.
• We cannot disprove the inaccessibility acocunt of forgetting: when discussing forgetting we do not dissassociate unavailable and inaccessible memories.
• Therefore, this theory is non-falsifiable.

• Parker et al (2006) - People who are unable to forget anything. This syndrome is characterised as  'nonstop, uncontrollable and automatic'.
• Parker tested patient AJ who kept a diary for 24 years (10-34). Researchers would pick a day at random and ask AJ questions about specific dates. She had incredibly accurate recall - it could be verified with her diary. 
• Ally et al (2012) - described patient HK with near perfect autobiographical memory. Observed hypertrophy (enlargement by 20%) of the amygdala. The amygdala is linked to emotion - emotional memories are recalled better as more important. Ally et al. suggests that memories were given greater emotional importance (due to a more active amygdala) and lead to improved memory. 

• Information is a form of updating (Bjork, 1978) - cognitive efficiency/economy. 
• Nor is forgetting a random process but shaped to our present needs - things that are no longer used / things that did not work / embarrassing/upsetting events / information that has changed. 

• Ebbinghaus (1885-1913) - learned a large number of 13-digit sequences. After interval (0s - 1 month), he tested memory for that sequence. He then measured the amount of time needed to relearn that sequence after the interval. This is known as the 'Ebbinghaus Forgetting Curve'.
• Bahrick (1984) also investigated the rate of forgetting. He used annual University reunions to test alumni's memory for foreign words - again, it was found that rapid forgetting was followed by stabilisation. This was demonstrated by a logarithimic curve. 
• Following stabilisation, Bahrick (1984) described these memories as within a 'permastore'. A permastore means that a memory is never lost/stabilised. Stays relatively permanent for a series of time. 

1) Trace Decay

2) Cue-dependent Forgetting

3) Interference

4) Consolidation

• Memories fade over time - connections weaken and then causes degradation. Items must be kept in the loop or else they will fade.
• This has been shown in animal models (Bailey and Chen, 1989), and in models of working memory, in association with the phonological loop (Baddeley and Hitch, 1974).
• This is known as the 'Word Length Effect' (Baddeley et al, 1975) -  whereby more time elapses for longer words before rehearsal, and therefore longer words should be forgotten to a greater extent.
• Decay is how we lose trivial memories encoded during the day (often during sleep) - Hardt et al, 2013.

" The notion that memory fades over time without an additional identifiable causal agent. The decay notion must assume a compensatory process such as rehearsal whenever forgetting over time is absent."

• Baddeley et al (1975) compared memory for sequences of 5 short words (e.g. wicket, phallic, ember. Avg articulation time of each word = 0.46s) with sequences of 5 long words (e.g. morphine, voodoo, humane. Avg articulation time = 0.77s). These words were matched for the number of syllables (all dysyllabic) and matched for the frequency in which they appeared in the english language. 
• Controlled for the level of interference (or as much as possible) and the only variable that altered was the articulation time. 
• Since there was a longer interval between presentation and recall for the longer words (lengthened retention interval), it is argued that those longer words are recalled less well because they have decayed more in memory. 

• Lovatt (2000) - argued that the word length effect was stimuli dependent. Using two sets of different words (matched for number of syllables, frequenct and phonological similarity) they initally reported that memory was better for longer words (experiment 1) and then found no effect of word length (experiment 2). When they used the same words as employed by Baddeley et al. (1975), they reported better memory for the longer words! (Experiment 3). 
• If time-based decay determines forgetting, we should consistently report the word length effect (as long as the words are matched appropriately). This therefore contradicts time-based decay in memory.
• Lewandowsky and Oberauer (2008) argue that articulation duration is quite complex cognitively, involving lots of different variables. Articulation duration is not simply a manipulation of time. It is difficult to increase time without interference also arising. 

• A failure to retrieve as the appropriate cue has not been presented. 
• Memory exists but issues in accessing that memory. 
• Meeter et al (2005) examined the forgetting curve for newsworthy events. Performance was better for multiple choice questions (52%) compared to free recall (31%). The MCQ answer options operated as a cue.
• Tulving and Pearlstone (1966) - found that participants recalled twice as many words when given a cue compared to free recall. Memory exists but requires the appropriate cue. However, performance not at 100%, so other explanations may be needed.
• Cues are important in memory as we do not learn information in a vacuum but also encode in the learning 'context'. This is known as Tulving's 'Encoding Specificity Principle' (1979), whereby retrieval is determined by the similarity between information available in memory and information available at retrieval. This is supported by context or 'state-dependent' studies. 

Memory is better when the context at learning and retrieval is the same. Context acts as a cue to recall. Examples include:

• Alcohol - Goodwin et al (1969)
• Diving - Godden and Baddeley (1975)
• Exercise - Miles and Hardman (1998)
• Chewing gum - Rickman et al (2013)

Cue dependent forgetting might also describe the effect of time on retrieval in which it may take longer to access a cue in the wrong context. 

• Our ability to remember current information is disrupted by previous learning and future learning. (Effects are not simply due to the passage of time).
• Highly typical events occur frequently and therefore increase over time. Interference may explain the role of time and apparent decay in forgetting. Crucially though, it does not always prefict forgetting as time increases.
• Competition assumption - Anderson et al (2004): Numerous events are attached to a single retrieval cue. The cue activates all linked memories and they compete. Reduced likelihood of selecting the correct memory. If too many items associated there is cue overload, and the cue ceases to be of utility.
• There are two main types of interference - retroactive and proactive.

Where there is a negative effect of new material (new memories) on the (older) to-be remembered stimuli. 

Testing retroactive interference:

• Learning phase (tree, chair, dog) -> 8 minute retention interval task: Unfilled / Attentional Task / Memory Task -> Immediate free-recall of the list. 
• Both attentional task and memory task both led to poorer recall relative to (unfilled) control (DeWar et al., 2007). 
• Time was kept constant across both conditions - RI not just due to memory demands.

Negative effects of prior learning (old memories) on the to-be remembered stimuli.

Testing proactive interference

• Recent probes task - Monsell, 1978.
• Does not explain why the rate of forgetting (Ebbinghaus) slows over time, as more interference should result in accelerated forgetting.
• A lot of overlap with other accounts.

• Memories are forgotten because they have not be adaquately consolidated (stabilised) in memory. 
• There are two forms of consolidation - systemic and synaptic.

'Rapidly formed and relatively long-lasting increase in the probablility that postsynaptic neurons in the hippocampus will fire in response to neurotransmitters released from the presynaptic neurons' - Wixted, 2004.

• This long-term enhancement in the signal between neurons occurs after a few hours/days.
• Disruption of this process leads to forgetting.
• Exposure to new memories (overlap with retroactive interference) will disrupt synaptic consolidation.
• Less forgetting when the retention interval is filled with sleep rather than wakefulness - Jenkins and Dallenbach, 1924. REM sleep with lucid dreams however, does not protect from forgetting - Phihal and Born, 1997.
• Less forgetting of material learnt immediately prior to intoxication - Bruce and Phil, 1997.

• Memories become independent of the hippocampus and are stored in the surrounding neocortex - McCaugh, 2000.
• Not susceptible to interference from other new memories (RI) as stored separately.
• Activation for recent memories freater for the hippocampus, whereas activation for older memories is greater for cortical memories - McKenzie and Eichenbaum, 2011.
• Evidence: Time-graded decline in memory for retrograde amnesia, Zola-Morgan et al, 1983.