lecture 18 - memory and EEG Flashcards
(24 cards)
overview
- Models of Recognition Memory
- Signal detection theory as a model of recognition
memory - Dual-process models of recognition memory
- How do we know that there are two processes that
contribute to memory judgements?
EEG - ERP Old/New effects: - The left-parietal old/new effect
- The mid-frontal old/new effect
recognition memory
we use it to determine whether we have seen something before or not and the context we can remember it
Memory
Need to make a decision whether we have encountered
something before or not (old/new) and sometimes the
context (e.g. left/right).
Encoding/study phase
Left/Right?
SAD
DOG
HAT
Retrieval/test phase
Left/Right/New?
BOOK
DOG
WAR
New items can be called:
distractors, lures, or foils.
ptps will see a word on the screen and have to do some encoding task with that word eg a shallow encoding task and ask if the word appeared on the right or left side of the screen or a deeper encoding task and ask is it a living or non living item that accesses the semantics = this is the study phase
there is then an interval
then we test peoples memory = retrieval or test phase
here we present words that ptpts saw in the encoding phase of the test and randomly intermix them with new words they didnt see in study phase. word is presented in centre of screen and ptps are asked if word is old or new or could ask if it appeared on right or left side of screen. these new items can be called distracters or foils
can be difficult to test memory ability as if pressed old word every time even for new words would get 100% correct for old items - so when we look at memory, we need to look at peoples ability to recognise whether something is old or not but also peoples ability to tell whether something is new or not and be able to tell when people are guessing - the signal detection model is very useful for this
signal detection model - graph in notes
Item is OLD Item is NEW
Response
OLD. Hit False Alarm
Response
NEW Miss Correct Rejection
the signal detection model derives from auditory perception - its the idea that when we are listening out for something we need to detect when something is there vs when somethings not
the model can help us understand how good peoples memories are
We need to know how many times they said an old item was old and how many times A new item was new and we’re gonna need to do something with that information and also going to need to take into account guessing so signal detection theory give us the tools to do this and we can separate true memory from guessing.
if item is old and ptp says is old = a hit
if item is old and ptp says its new = miss
if item is new and ptp says its old = false alarm as a false memory
if item is new but you says its new = a correct rejection
Signal detection model
- Memory (or familiarity) strength is
a continuous variable that varies from high to low - plotted on the x axis on the graph - right side is high strength - proposes memory traces have strength values and that reflects their activation in memory and the level of familiarity
- memory traces vary in strength eg if item repeated to you lots of time it should be associated with higher memory strength
- we can plot the distribution of memory strength for old items and new items
- Distribution assumed to be normal
and to overlap. - old items should have more memory strength so should be higher up the scale than new items. New items also have some memory strength associated with them we would anticipate that to be lower than the old items this could be because they’re similar to some of the old items. - there is a distribution for new items and a distribution for old items
- Response criterion. - we place this on the distribution of memory strengths and if any items comes along which has a memory strength to the right of the response criterion line we say its old but if its to the left its new
- Discrimination sensitivity e.g. d’. - the ability to discriminate between old and new items. it’s measured from the average of the old distribution to the average of the new distribution. the further apart those two distributions are the less overlap there is and the easier it is to discriminate old from new items and vice versa. if old and new words are similar the distributions would be very close
- Different criteria e.g. conservative
and liberal. - - you can move the response criterion eg if asked ptps to only give old resposne if they are absolutely sure the item is old you would move the response criterion closer to the right side as if you want to make sure that the items are old then you’re only going to select items which are very high memory strength in order to make an old response to them = a more conservative criterion - more lax criterion would go down the other end - the left
the items on the graph that end up just on the right side of the repose criterion are new items that you are staying are old which is a false alarm
signal detection theory is a useful way of looking at memory
Dual-process accounts
Two processes contribute to recognition memory
judgements: how familiar a stimulus seems (Familiarity)
or by recalling the particulars of the experience
(Recollection) – e.g. Yonelinas, 1999.
- Familiarity: awareness of a prior encounter but with no
recovery of contextual details. Fast and automatic.
Modelled in the same way as the signal detection model. - Recollection: recovery of contextual details e.g. location,
colour, task….Slow, more attention demanding process.
What is the evidence that there are two
process that contribute to memory
judgements?
Focus on EEG and ERP evidence
Electroencephalography (EEG)
A non-invasive technique which records the electrical activity of the brain.
involves placing a cap on patients head and electrodes are placed on the cap which measure certain parts of the scalp - the cap allows consistency of placement across ptps. the electrodes record the electrical activity in the brain. each line in graph is activity detected in one electrode and it gets longer as time goes on
The strength of EEG is that it has
excellent temporal resolution. - we can tell to the millisecond what is happening in the brain. fMRI its bad
spatial resolution - if you want to find out where in brain something is happening - on EEG is not good. fMRI is good
one way to look at brain activity is to look at the frequency of brain activity eg
beta wave - concentrating
alpha - relaxed
delta
Event-Related Potentials (ERPs)
a technique that means you can understand brain activity that is elicited by certain cognitive tasks by certain stimuli or events we are interested in
ptps have cap and electrodes on
A. Show participants old and new items
randomly intermixed and record their
response, while recording EEG.
trigger are then sent from the stimulus PC to the EEG computer which marks when a stimulus appears on screen - when old and new items appear
B. Epoch (cut-up) the EEG time-locked to
the onset of the item. Pre-process data
e.g. exclude artifacts, filter…
C. Average together trials of
the same type e.g. hits,
correct rejections
the idea behind ERPs is that activity that is elicited by that stimulus will be time locked to it so that signal will get stronger when we average together lots of trials of the same type - any random noise will cancel out so we should be left with a signal that we see in repose top the item
the more trials addd together the better the signal
ERP evidence for two memory
processes
Relevant findings come from analyses of old/new effects.
These are the differences in neural activity elicited by
old (previously studied) and new (previously unstudied)
items that attract correct judgments.
looking at hits and then correct rejections are acting as a baseline/control - the idea is that when you see a new item and do a correct rejection you should elicit the same processes as when you see the old item but critical thing with the correct rejection if you’re not going to get anything back from memory, so we see the same processes except the critical memory processes so acts as a control we measure the hits against
study
*
DOG
| |
“Concrete
or Abstract”
Test
*
DOG
| |
“Old or New”
in the retrieval phase we time lock onto the onset of the item and split these items into categories of response things like hits and correct responses
- The logic of this contrast is that a reliable ERP difference
between these items are electrophysiological indices of
processes that reflect or are contingent on successful
memory retrieval.
hits should indicate successful memory retrieval
old/new effects
There are a family of old/new effects
that are distinguishable on the basis
of their time course, scalp distribution
and sensitivity to experimental
variables.
* The left-parietal ERP old/new effect - see on left side of scalp and towards back
* The mid-frontal ERP old/new effect - front and middle of head
The left-parietal ERP old/new effect
and source tasks
Wilding et al. (1995)
* At study words presented auditorily or visually.
At test: old or new ? if said item was old they were asked for the modality In which they encountered it in the study phase
|
Auditory or Visual at study?
This ERP was larger when participants make a correct
source judgments (when got modality correct) compared to incorrect after a hit.
waveform in notes - its from a left parietal site
left is the onset of the item in test phase and each line corresponds to the average - top line = source correct
- dotted line - source incorrect
- thick black line - correct rejection
correct rejection is baseline and there is a big difference between that and source correct but theres not much difference between source incorrect and correct rejections
Which process might this ERP be indexing (recollection or
familiarity)? its recollection which is the process whereby if you want to retrieve the contextual details - should be present when you can correctly remember the source details and be absent when you cannot successfully remember the source details. as there is an index for source correct but not one for source incorrect this indicates that this is an index of recollection
The left-parietal ERP old/new effect
The effect:
- onsets around 400-500 ms post-stimulus
- often largest at left-parietal scalp locations
- typically has a duration of 500-800 ms post-stimulus
Seems to be
an index of
Recollection
Taken from: Wilding and Ranganath (2012)
The mid-frontal ERP old/new
effect and criterion
Azimian-Faridani & Wilding (2006) studied the effects of
adopting a conservative or liberal decision criteria on the
mid-frontal effect.
conservative: emphasising old responses, give an old
response only when confident. - the memory strength would fall above the c line on graph
liberal: emphasising new responses, give a new response
only when confident. - strength would be other side of L line so dot catch any old items
If the mid-frontal ERP
old/new effect is an index of
familiarity will the index be
larger in the conservative or
the liberal condition? we would expect it to be larger in the conservative condition - only the items which have the very highest memory strength will be in the conservative condition
if we average the memory strength of these items its going to be higher than if we average the memory strength of the items in the liberal conditions as more Lower memory strength items are in that condition
results - hits in the conservative
(C) condition were larger than
those in the liberal (L) so they are more positive going. graphs in notes
this indicates its an index of familalirty as the familiarity index should vary according to where we place our repose criterion
The mid-frontal ERP old/new
effect
The effect is:
- evident between 300-500 ms
- largest at midline frontal sites
- comprises a relatively greater positivity for hits compared
to correct rejections.
Taken from: Wilding and Ranganath (2012)
Seems to be
an index of
Familiarity
Yu and Rugg (2010)
Aim: to electrophysiologically dissociate the neural
correlates of Recollection and Familiarity
.
Modified Remember/Know paradigm at test
Remember/Know procedure
- One method to isolate the contributions of recollection
and familiarity (Tulving, 1985). - For each judgment in the test phase participants are
asked why they feel they recognise the item - if you recognise that you saw the item in the study phase you need to remember something else from the encounter : - Remember: they consciously recollect the particulars of
the study event. - Know: they feel that they have seen the item before but
there is no memory for the details of the event. - New: the item has not been presented in the study
phase
Yu and Rugg (2010)
Modified Remember/Know paradigm at test: Remember,
confident old (Know), unconfident old (Know but don’t remember details), unconfident
new, confident new
- Prediction for the mid-frontal old/new effect? indexing memory strength or familiarity
Should covary with recognition confidence i.e. CO > UO >
UN > CN - each one should be more positive than the one after - Prediction for the left-parietal old/new effect? a index of recollection
Will be elicited by Remember items, but not either old
class of item
findings in notes
mid-frontal old/new effect - there is a graded effect - its more positive going for confident old compared to unconfident old and compared to unconfident new
left-parietal old/new effect - you get a bigger left parietal effect for the recollect response compared to the confident old response
Summary
- There are two temporally, topographically and functionally distinct
old/new effects in the electrical record. - The mid-frontal ERP old/new effect happens between 300-
500 ms and the left-parietal ERP old/new effect happens between 500-
800ms. - The mid-frontal ERP old/new effect is seen at mid-frontal
scalp locations and the left-parietal ERP old/new effect appears at left-
parietal scalp locations. - The mid-frontal ERP old/new effect seems to be a putative
index of Familiarity and the left-parietal ERP old/new effect seems to be
an index of Recollection. - This strongly supports the dual-process model, that two processes
contribute to memory judgements
recognition memory
Recognition Memory
1. What is Recognition Memory?
Recognition memory involves identifying whether a stimulus has been encountered before.
Unlike recall, where we generate information from memory, recognition requires us to judge whether we’ve seen something before.
Examples include:
Recognizing a phone number on a list.
Identifying someone on the street.
Identifying a perpetrator in a police lineup.
2. Recognition vs. Recall
Recall: Recalling or generating information without being presented with the original stimulus.
Recognition: Making a decision by seeing the intact stimulus and determining if it was encountered before.
3. Discrimination in Recognition Memory
Key aspect: In recognition, we must discriminate between “old” (previously encountered) and “new” stimuli.
Test Types:
Forced-choice recognition test: Old and new items are presented together, and the person must choose which one they’ve encountered before.
Yes/No recognition test: One item at a time is shown, and the person must decide whether it’s old or new (with items intermixed).
4. Distractors (Lures or Foils)
These are new stimuli included in recognition tests to evaluate how well a person can discriminate.
They help assess whether someone’s recognition judgments are accurate, as a person must distinguish between old items (experienced) and distractors (new).
5. Measuring Recognition: The Error Dilemma
Recognition errors: A single error does not necessarily reflect poor memory. Even individuals with good memory can make mistakes.
How to interpret mistakes?: How do we assess recognition accuracy when a person has made errors?
If a person correctly identifies 85% of old items but makes 10% mistakes, is their memory worse than someone who only identifies 40% of old items but has only 5% mistakes?
6. Guessing and Biases
Guessing plays a large role in recognition memory:
Uncertainty about whether an item is “old” or “new” may lead to a guess.
In police lineups, for example, witnesses may guess based on who feels familiar, even if they don’t have strong evidence.
Influence of guessing:
Person A: No penalty for incorrect guesses—more likely to guess “Yes” even to new items.
Person B: Harsh penalties for mistakes—more likely to be conservative, leading to fewer “Yes” answers, even when uncertain.
7. Memory vs. Decision Making
Measuring recognition memory is not just about assessing memory, but also about separating judgment biases.
There needs to be a method that distinguishes between genuine memory and decision-making biases.
8. Key Concept Summary
Key Concept Explanation
Recognition Memory Identifying whether you’ve encountered a stimulus before.
Discrimination Judging “old” (previously encountered) vs. “new” (new) items.
Forced-Choice Recognition Test Old and new items are presented together, and the participant must choose.
Yes/No Recognition Test One item at a time is presented, and the person must decide whether it’s old or new.
Distractors (Lures/Foils) Non-studied items included in tests to assess accuracy.
Guessing Bias Uncertainty about whether an item is “old” or “new” leads to guesswork.
Judgment Biases in Recognition The influence of guessing tendencies can affect recognition judgments.
9. Practical Implications
🧠 Recognition memory is not just about what you remember but also how you decide what you know.
In real-life situations (like eyewitness identification), guessing and biases can significantly influence the accuracy of recognition memory.
Testing and assessing recognition must consider how much decision-making influences the response, not just the memory itself.
signal detection theory as a model of recognition memory
Signal Detection Theory (SDT) as a Model of Recognition Memory
1. Origins of Signal Detection Theory
Signal Detection Theory (SDT) originated in auditory perception research (Green & Swets, 1966).
In auditory experiments, people listen for a faint tone against background noise and must decide whether or not they heard the tone.
Four possible outcomes:
Hit: Correctly detect a tone.
Miss: Fail to detect a tone when it’s present.
False Alarm: Incorrectly report hearing a tone when there’s none.
Correct Rejection: Correctly report no tone when none is presented.
2. Recognition Memory as a Similar Process
Recognition tests share a similar structure to the auditory detection task.
Participants need to decide if a stimulus seems familiar enough to be classified as “Old” (previously studied).
Four possible outcomes in recognition:
Hit: Correctly identify a studied item as “Old.”
Miss: Fail to identify a studied item as “Old.”
False Alarm: Incorrectly classify a new item as “Old.”
Correct Rejection: Correctly classify a new item as “New.”
3. Signal Detection Theory in Recognition Memory
Memory traces are thought to have different strength values that reflect their activation in memory.
Stronger traces: Items studied more attentively or repeatedly.
Weaker traces: New items, or those encountered less frequently.
Familiarity of items:
Both old and new items can feel familiar, but typically, old items feel more familiar due to recent exposure.
New items may feel familiar if they resemble studied items or have been encountered outside the experiment (e.g., in everyday life or media).
4. The Familiarity Continuum & Distributions
Familiarity distributions:
The familiarity of old and new items follows different normal distributions.
The average familiarity for old items is usually higher than for new items (due to prior exposure).
However, overlap between the distributions can occur, as:
Poorly encoded old items may have low familiarity.
New items may feel unexpectedly familiar (due to similarity to studied items or prior experiences).
5. Discriminating Between Old and New Items
The distance between the averages of the old and new distributions is key to recognizing whether a memory is old or new.
This distance is represented by d’ (“d prime”), which quantifies the ability to discriminate between old and new items.
6. The Criterion for Judging “Old”
Recognition Judgment: A person sets a criterion level of familiarity:
Above the criterion: Item classified as “Old.”
Below the criterion: Item classified as “New.”
This criterion can be adjusted (shifted left or right) based on the person’s judgment strategy.
7. Liberal vs. Conservative Judging Strategies
Liberal Strategy:
Lower the criterion (shift left), meaning more items are judged as “Old.”
Results: More hits, but more false alarms (judging new items as old).
Conservative Strategy:
Raise the criterion (shift right), meaning fewer items are judged as “Old.”
Results: Fewer false alarms, but more misses (failing to recognize old items).
Balanced (Unbiased) Strategy:
Criterion placed between the means of the old and new distributions, leading to balanced hits and false alarms.
Beta (β) represents the tendency to guess, indicating how strict or lenient a person is when judging “Old.”
8. Mathematical Tools in SDT for Recognition Memory
By calculating the hit rate (correctly identifying old items) and the false alarm rate (misidentifying new items as old), researchers can compute:
d’: Discrimination ability.
β (Beta): Tendency to guess.
This allows researchers to separate memory strength from judgment biases.
- Challenges to Signal Detection Theory
The Word Frequency Effect:
Typically, high-frequency words should be better recognized (due to stronger memory traces from repeated exposure).
However, low-frequency words are often better recognized, a phenomenon known as the word frequency effect in recognition memory.
SDT struggles to explain this, suggesting that additional factors (other than just memory strength) influence recognition.
Key Concept Summary
Concept Explanation
Signal Detection Theory (SDT) A framework for understanding recognition memory, based on auditory detection tasks.
d’ (d prime) Measures the ability to discriminate between old and new items.
Beta (β) Measures a person’s tendency to guess when making a recognition judgment.
Criterion for “Old” Judgment A threshold of familiarity set by the person to decide if an item is “Old” or “New.”
Liberal Strategy Lowering the criterion, increasing hits but also false alarms.
Conservative Strategy Raising the criterion, decreasing false alarms but increasing misses.
Word Frequency Effect The phenomenon where low-frequency words are better recognized than high-frequency words.
Practical Implications
🧠 Signal detection theory offers a comprehensive framework for understanding recognition memory. It emphasizes that recognition judgments are not solely based on memory strength but also influenced by decision-making processes (e.g., guesswork, bias).
Practical significance: This theory is useful in fields like eyewitness testimony, where guessing can impact the accuracy of identifications
dual-process accounts of recognition of memory
Introduction to the Dual-Process Model
Recognition memory is proposed to involve two distinct processes: familiarity and recollection.
Familiarity-based recognition: You recognize something because it seems familiar, but you cannot recall specific details about the experience.
Recollection-based recognition: You recognize something by retrieving specific details (e.g., context, source, or time) about the experience.
2. The Role of Familiarity and Recollection
Familiarity is described as fast, automatic, and effortless, generating a sense of memory strength without recalling specifics. This process is well-described by signal detection theory.
Recollection is slower, controlled, and attention-demanding. It involves retrieving context or specific details related to the experience, much like cued recall.
3. The Remember/Know Procedure
Remember/Know Procedure (Tulving, 1985):
Participants are asked whether they recognize an item based on:
“Remember”: Consciously recalling specifics of the event.
“Know”: A sense of familiarity without recalling details.
“Remember” responses measure recollection.
“Know” responses measure familiarity.
Criticism: Some researchers argue this method reflects a single signal detection process based on familiarity, rather than distinct processes (e.g., Rotello & Zheng, 2008).
4. The Process Dissociation Procedure (PDP)
PDP (Jacoby, 1991): A method to isolate familiarity and recollection by creating two conditions:
Inclusion condition: Say “Yes” to any item (seen or heard).
Exclusion condition: Say “Yes” only to items that were heard.
Estimating Familiarity: Errors in the exclusion condition (e.g., saying “Yes” to an item seen, not heard) suggest familiarity without recollection.
By subtracting these errors, researchers can isolate recollection.
5. Evidence Supporting Two Processes
Distraction and Attention:
Recollection is more sensitive to distraction and requires attention. If attention is divided during encoding, recollection later will suffer, but familiarity can persist.
This suggests that familiarity is an automatic process, while recollection is controlled and attention-dependent.
Age and Prefrontal Cortex:
Older adults and individuals with prefrontal cortex damage often exhibit impaired recollection but intact familiarity. This supports the view that recollection is more demanding.
6. Neuroanatomical Support for Dual-Process Models
Familiarity and Recollection rely on distinct brain structures:
Familiarity is linked to the perirhinal cortex, adjacent to the hippocampus.
Recollection relies heavily on the hippocampus, particularly the posterior hippocampus.
Jon’s Case (Developmental Amnesia):
Jon, a patient with hippocampal damage, shows impaired recollection but retains a normal sense of familiarity. His intact perirhinal cortex accounts for this.
Neuroimaging Evidence:
Studies using the Remember/Know procedure show that hippocampal activation is higher for recollection (e.g., remembering specific details), while perirhinal cortex activation is associated with familiarity (e.g., a general sense of knowing).
7. Key Neuroimaging Findings
Hippocampal Activation:
Increased activation in the hippocampus occurs when participants “remember” specific details about an event.
Perirhinal Cortex Activation:
Activation in the perirhinal cortex is associated with familiarity judgments, reflecting the automatic sense of knowing.
8. Practical Implications
🧠 The dual-process theory of recognition memory emphasizes that recognition is not just about how familiar something feels. Instead, it involves two distinct retrieval processes:
Familiarity: Fast, automatic, and linked to the perirhinal cortex.
Recollection: Slow, controlled, and tied to the hippocampus.
Table Summary
Process Characteristics Brain Region
Familiarity Fast, automatic, based on memory strength. Perirhinal Cortex
Recollection Slow, controlled, involves recalling event details. Hippocampus (posterior)
9. Conclusion
Dual-process models suggest that recognition memory arises from two independent processes—familiarity and recollection—each supported by different neural structures.
These processes interact to provide a full picture of recognition: one based on memory strength (familiarity), the other on contextual details (recollection).
source monitoring
- Introduction to Source Monitoring
Source monitoring refers to the process of identifying the origin or source of a retrieved memory. For example, asking questions like:
Did I hear this story from Susan or Maria?
Did I park here today or last week?
Did I read this fact in the National Enquirer or Consumer Reports?
Source monitoring helps distinguish whether an event or piece of information originated from personal experience, external sources, or imagination. - Source Monitoring as a Post-Retrieval Process
Source monitoring involves post-retrieval monitoring, which means it occurs after a memory trace has been activated.
This process relies on controlled cognitive processes, primarily mediated by the prefrontal cortex (Mitchell & Johnson, 2009; Spaniol et al., 2009). - Source Misattribution Errors
Source misattribution occurs when a person mistakenly attributes a recollection to the wrong source. Common examples:
Confusing which friend told you something.
Forgetting whether you learned a fact through reading or hearing it.
These errors can happen in casual conversations where individuals don’t make an effort to recall the source accurately.
Example: Confusing which grandchild likes which hobby or which favorite joke was told by whom. - How Source Monitoring Works
People monitor the sources of their memories by recalling contextual details of the original experience.
These details allow individuals to determine the origin of a memory based on certain regularities:
Auditory details indicate that something was heard.
Visual details indicate that something was read or seen.
The perceptual richness of the memory helps determine whether it was a real event or something imagined. - Reality Monitoring
Reality monitoring refers to the process of distinguishing between memories of real events and those that were imagined.
Example: Evaluating whether we actually saw a picture or just imagined it.
Mistakes in reality monitoring occur when people mistake imagined details for perceptual experiences:
Example: If someone forms a mental image of a word, they may later mistakenly believe they saw an actual picture of the object (Henkel, Franklin, & Johnson, 2000).
Hallucinations in conditions like schizophrenia arise when individuals can’t distinguish their imagined experiences from real ones, due to faulty reality monitoring. - Cognitive and Neural Mechanisms
The ability to distinguish between real and imagined memories relies on activity in the anterior prefrontal cortex and its structural integrity (Simons et al., 2017).
Individuals with impaired prefrontal cortex function may struggle with reality monitoring, leading to more frequent misattributions. - Source Monitoring in Psychologically Healthy People
Psychologically healthy individuals also vary in their ability to distinguish between real and imagined events. Some people are better than others at accurately monitoring the source of their memories.
Good reality monitoring is important for accurate memory retrieval, while deficits can lead to confusion and misremembered details. - Applications and Future Discussions
Source monitoring errors play a significant role in areas like:
Eyewitness memory (Chapter 12).
Motivated forgetting (Chapter 10).
Table Summary
Aspect Details
Source Monitoring Identifying the origin of a memory (e.g., “Did I hear this from Susan or Maria?”)
Reality Monitoring Distinguishing between real events and imagined ones (e.g., “Did I see this picture or just imagine it?”)
Source Misattribution Mistakenly attributing a memory to the wrong source (e.g., confusing friends or events)
Cognitive Mechanism Controlled by the prefrontal cortex; errors can occur in casual or distracted contexts
Neural Basis Involves the anterior prefrontal cortex; damage can impair source/reality monitoring - Conclusion
Source monitoring is a crucial cognitive process that helps us evaluate and identify the origin of memories. While generally effective, this system is prone to misattributions, especially in casual or less attentively monitored situations.
Understanding source misattributions and reality monitoring is important for various areas of psychology, including eyewitness testimony and psychopathology
Wilding, E. L. & Ranganath, C. (2012).
Electrophysiological correlates of episodic memory
processes. In: Luck, S. and Kappenman, E.
eds. The Oxford Handbook of Event-Related
Potential Components. Oxford: Oxford University
Press, pp. 373-396.
Overview of ERP Studies and Recognition Memory
ERP (Event-Related Potential) studies focus on recognition memory, which involves identifying whether a stimulus (e.g., a face) has been encountered before.
There are two major theoretical perspectives on how recognition occurs:
Signal Detection Models: Recognition is seen as a process of assessing the memory strength of a stimulus, where previously encountered items are remembered more strongly.
Dual-Process Models: Recognition relies on two separate processes—recollection (recovering qualitative details from a prior episode) and familiarity (judging the strength of memory without detailed context).
Old-New ERP Effects
Old-New Effects are differences in brain activity when recognizing old (studied) versus new (unstudied) items. These effects are crucial in understanding recognition memory.
Early ERP studies found a posterior midline maximum modulation (more positive for old items) that appeared when individuals made correct recognition judgments. This suggests that the effect is linked to true memory recognition, not just familiarity or prior exposure.
Left-Parietal ERP Old-New Effect
This old-new effect is typically strongest at the left-parietal region of the brain, observed between 500-800 ms after stimulus presentation.
It is often linked to recollection because:
It correlates with accurate contextual judgments (e.g., remembering the modality in which an item was presented).
The effect is reduced in individuals with brain damage or cognitive impairments affecting recollection.
Studies support the link between this left-parietal ERP effect and recollection through behavioral and neural data.
Midfrontal ERP Old-New Effect
A second ERP old-new effect appears between 300-500 ms, strongest at frontal midline sites, and is thought to be related to familiarity.
Studies show that incorrect old responses to similar lures (words similar to studied items) elicit this midfrontal effect, which suggests it corresponds to familiarity rather than recollection.
The dual-process models propose that familiarity is a gradual process, while recollection is more of a threshold process.
Support for Dual-Process Models
The two distinct ERP effects—the left-parietal effect (linked to recollection) and the midfrontal effect (linked to familiarity)—provide support for dual-process models.
Double dissociation findings (where changes in one effect occur without changes in the other) further strengthen this view, as shown in studies where:
Recollection (indexed by the left-parietal effect) is linked to “remember” responses, while familiarity (midfrontal effect) correlates with confidence judgments.
Conclusion
ERP studies support the idea that recognition memory is based on two distinct processes: recollection (linked to the left-parietal old-new effect) and familiarity (linked to the midfrontal old-new effect).
These findings align with dual-process models, suggesting that recollection operates as a threshold-like process and familiarity as a graded strength measure
