Lecture 15 - Recognition Memory & EEG Flashcards
What is a recognition memory test?
- See objects and some have been encountered before and other have not.
- Which object have you encountered before?
- Also need to recognise some of the details of the event e.g in study phase, show ppt words on a screen - gap - memory testing after through words in study phase and random words too
What is the signal detection model?
- Memory is a continuous variable
- Distribution is assumed to be normal and overlap (depends on manipulation of tasks)
- Response criterion: black line on graph. If it's on the right, people will say it is old information regardless, and vice versa
- Discrimination sensitivity: is distance between two peaks: discrimination from old to new information
- Different criteria
What was a notation that shows signal detection?
- Results: if item and response is old = hit, if response old and it is new = false alarm
- If response is new & item is old = miss, if item is new = correct rejection
What are the dual process accounts?
- Familiarity: awareness of a prior encounter but with no recovery of contextual details
- Fast and automatic
- Modelled same as signal detection theory
- Recollection: recovery of contextual details e.g location.
- Slow & more attention demanding
What is the evidence that are two processes that contribute to memory judgements?
- EEG & ERP evidence
What is EEG?
- Electroencephalography
- Non-invasive that records electrical activity of the brain via oscillations
- Strength = excellent temporal resolution e.g to the millisecond
- Not very good at spatial resolution
What was a study conducted for ERPs?
- Ppts were shown old and new items randomly intermixed and record their response whilst recording EEG (after study phase)
- Cut up EEG time locked to the onset of the item
- Average together trails of the same type e.g all the hits = memory signal should get stronger and noise will drop out
What is ERP evidence for two memory processes?
- Relevant findings come from analyses of old/new effects
- Correct judgement with new word = baseline as non-memory is tested
- Contrast correct judgements to old and new (to test memory vs non-memory)
- If reliable ERP diff found = indicates we have an index of processes that reflect successful memory retrieval.
What is the left-parietal ERP old/new effect?
- Onsets 400-500ms post-stimulus
- Largest at left-parietal scalp locations
- Duration of 500-800ms
- Index of recollection
How does the left parietal old/new effect show recollection experimentally?
- Words presented visually or auditorily in study phase
- Tested all visually presented Ppts: ask if word is old/new & ask if heard/seen (source)
- ERP was larger when ppt made a correct source judgment compared to incorrect
- Shows recollection as source correct shows they remembered the context (source of how they received info)
What is the mid-frontal ERP old/new effect and criterion?
- Studied effects of adopting tory/liberal decision criteria on mid-frontal effect
- Tory: asked to give old response when confident item is old
- Liberal: asked to new response only when confident
- Both criterions get pushed in opposite directions
- Index of familiarity
- Hits in tory condition were larger than liberal
What did Yu and Rugg do?
- Aimed to electrophysiologically dissociate the neural correlates of recollection and familiarity (Combine both studies above)
- Used modified Remember/Know paradigm as a test
- For each judgment in the test phase, ppts asked why they feel they recognise the item
- Remember: consciously recollect particulars of study event e.g word was in red/emotions (recollection)
- Know: Feel they have seen item before but no memory for details of event (familiarity)
- Know split into two: confident/unconfident
- New: item has not been presented in study phase
- Prediction for mid-frontal old/new effect: Should covary with recognition confidence
- Prediction for left-parietal old/new effect: Elicited by 'Remember' items but not either class of old
