Midterm 2- Study Guide.doc

Midterm 2- Study Guide Types of memory Semantic memory = memory for facts and general information about the world Organized into networks Nodes represent objects or concepts There are superordinate and subordinate nodes Basic levels = levels of the hierarchy that are easier to access than others Atypical exemplars = don't follow the generalization that the superordinate node describes (e.g. a penguin is a bird that can't fly) Episodic memory = memory for a specific event (what) that occurred ina specific time (when) and place (where) Relationship between semantic and episodic memory Theory 1: episodic memory grows out of semantic memroy Theory 2: semantic memory is information we've repeatedly encountered, so that specific learning encounters are blurred and only the semantic "fact" content remains Types of memory tests Free recall = listing anything you can remember Influenced by primacy and recency effects Cued recall = cues of the experience are presented E.g. stem completion Recognition = indicate if you've seen the particular word before Stages of LTM Encoding = placing information into memory Can be an incidental byproduct of the active stimulus processing carried out during an event Storage/consolidation = keeping the information in a permanent store Retrieval = bringing memories back to mind Episodic encoding = processes that yield a durable memory trace so that an event can be subsequently remembered Attention = facilitates encoding of specific stimulus feature Selective attention and visual search Attention as a spotlight Stimuli that are landing at the focus of attention are responded to more quickly and accurately Dual attention theory Goal-directed/top-down attention = when doing a task, we can strategically allocate attention to things Reactive/reflexive attention = allows attention to be captured by things that are unexpected and salient Posner's Task Arrow indicates where something is likely going to appear, so it's strategically to your advantage to willfully shift your attention Working memory in space interacts with oculomotor control systems Stimulus will either appear there or in an uncued position Bottom up form of attention Results: On valid trials, Ps are faster because they shift their attention On invalid trials, they're slower Gain modulation = visual attention potentiates the neural response that plays a role in perceiving stimuli increased sensitivity due to goal-directed attention --> increased performance and speed Primate study When visual objects land in an attended location, the neuron responds at a higher firing rate than when they land in unattended areas PET study There is enhanced response on the side they're attending Full vs. divided attention Attention during encoding IVs Full attention: intentional learning of visually presented words Divided attention: word learning + secondary task (monitoring auditory tones) Tones were either slow or fast Results Full attention > divided slow > divided fast Levels of processing Shallow vs. deep encoding: difficulties and implications Structural = (shallowest) surface perceptual features of the presented stimulus Phonological/lexical = convert stimuli to speech-based codes Semantic = (deepest) access and evaluate the meaning of the stimulus Experimental evidence 4 IVs: Intentionally commit words to memory Judge pleasantness of word (deep) Judge whether there is an 'e' in the word (shallow) Judge how many letters are in the word (shallow) Results Deep encoding is about twice as good as shallow encoding Intentional encoding is no better than incidental deep encoding Difficulties of LoP Circular logic If you encode something deeply you'll remember it better, but there's no evidence for deep encoding other than better remembering This is fixed by showing that depth of processing only works when you're recalling things in the same way Transfer appropriate processing = the processes that we're using when we're trying to retrieve a memory is similar to the processes we used when we formed the memory in the first place If you use the same processes during encoding and retrieval, it's more likely that you'll remember Experimental evidence Ps had to say whether a word fit a blank IVs Meaning condition Phonology (rhyming) condition 2 tests Asked if they saw a certain word before Asked if they saw a word that rhymed with something before Encoding and retrieval conditions had to match Semantic retrieval Retrieval attempts tend to be semantically-based Semantic processing may be more distinctive, decreasing interference effects Links events to other knowledge in semantic memory, providing more retrieval cues Relationship between levels of processing effects and attention Attention facilitates encoding of specific stimulus features LoP varies which features of something are effectively encoded Change blindness = we're blind to large scale perceptual changes when the stuff that change isn't what we were attending to Role of PFC in encoding PFC = regulates what enters the sensory cortices In charge of what gets emphasized, what gets encoded, what gets Top-down cognitive control PFC lesions Leads to modest episodic memory deficits, but not global amnesia Event content The content of an event (phonological, semantic, or visuospatial) determine what PFC-posterior neocortical circuit contributes to learning Power of retrieval Generation effect = people who are asked to generate words have better memory later than those who are just asked to read them Engages retrieval during studying Engages greater elaboration/processing of meaning/phonology Experimental evidence 2 conditions: synonym and rhyme Those who were asked to generate their own examples of synonyms and rhymes remembered the words better than those who just read them Judgements of learning vs. memory performance Testing Effect IVs SSSS SSST STTT Judgement of learning: SSSS > SSST > STTT Actual memory performance on an immediate test: SSSS > SSST > STTT Actual memory performance on a delayed test: STTT > SSST > SSSS Attention mediates the testing effect Better recognition and source memory with full vs. divided attention "Desirable difficulties" and encoding When you think about things more critically, the more likely it is that you'll remember them Distribution of practice Spacing effect = longer las between study presentations seem to give better memory than short ones Experimental evidence Presented paris of words IVs Massed: presenting the same pairs in succession Spaced: Presenting other pairs of items between repetitions of the target pair Results: Spaced practice is much better, especially when combined with the generation effect Deficient processing = during massed/short lag trials, the second occurrence of an item isn't processed fully Encoding variability = The longer apart the study events are, the more variable the encoding is because there are richer and more varied cues Stimulus sampling theory = Every time you study, you get different stimuli that can serve as retrieval cues The longer lags there are, the more different stimuli you have to serve as retrieval cues, and the more likely you'll be to remember Spacing/lag effects Expanding practice = allowing the spaces to get bigger is even better than spacing them out equally Neural bases of episodic memory Agnosia = relatively selective disruption in the ability to process a particular kind of information Associative visual Auditory Tactile Transient Global Amnesia = temporary disruption of memory Functional amnesia = amnesia that seems to result from psychological causes Infantile amnesia = forgetting events that happened during infancy Global amnesia = cuts across all modalities, all classes of stimuli, all ways of probing for memory Not being able to remember things- Any type of assessment (free recall, cued recall, yes/no recognition, multiple choice recognition, learning to criterion Caused by MTL damage Anterograde vs. retrograde amnesia Anterograde amnesia = the inability to remember new events Retrograde amnesia = the inability to recollect memories acquired before an injury After a severe brain injury, you usually see anterograde coupled with temporally graded retrograde MTL Material-specific MTL involvement Left hemisphere lesion: problems with verbal learning Right hemisphere lesion: problems with nonverbal (spatial, visual) learning MTL hierarchical structure Dorsal parietal = top-down attention Activation is positively correlated with remembering Ventral parietal = stimulus-driven Activation is negatively correlated with remembering Stimulus-driven attention can hinder encoding if it's captured by irrelevant things Relational Memory Theory = The MTL, in interaction with neocortical storage sites, supports conjunctive memory binding & activation, representing the relations among perceptually distinct items of scenes or events, and mediating their retrieval  Tenets (RAR DRIF) Declarative memory is fundamentally relational Represents and retrieves all manner of relations among the elements of events Hippocampus contributes to the critical process of relational memory binding and activations of conjunctions Hippocampus supports declarative memory Rapidly acquires representations Parahipocampal/perirhinal regions contribute to memory for individual items Representations are flexibly accessible Hippocampus vs. MTL cortex (perirhinal/entorhinal/parahippocampal) MTL cortex = gathers different bits of information and projects them onto the hippocampus Item recognition Hippocampus = convergence zone in which everything in neocortex gets projected in Associative Glues bits of memory together Conjunctive representations = capture conjunction of event elements that are occurring in a moment in time Functional distinctions within the MTL circuit Memory retrieval as prediction Prediction error/novelty detection; CA1 comparator hypothesis CA1 Comparator Hypothesis From our experience, we use certain cues to predict the future When our predictions don't match reality, CA1 lights up Subsequent forgetting paradigm = MTL is more active when things are subsequently remembered Memory consolidation Consolidation = processes that transform a memory trace into a durable representation that is not dependent on the MTL Ribot's law = new perishes before the old, the complex before the simple Consolidation period = length of time during which new memories are vulnerable and easily lost What is the evidence it occurs? Temporally graded retrograde amnesia = things that were experienced closer to an injury are more likely to be lost than things in the distant past Standard model vs. multiple trace theory Standard model 1. Conjunctive representation in the hippocampus 2. Gets replayed in neocortical areas 3. Eventually, the cortical representations bind together Memory = MTL pointer and neocortical traces 4. Eventually, the neocortex binds together the information without the help of MTL structures Multiple trace theory Episodic memories are encoded by an ensemble of hippocampal and cortical neurons, and the cortical neurons never, in normal circumstances, become fully independent of the hippocampal neurons Role of sleep in consolidation = replays things that were present during the day Consolidation during REM sleep (Rodent Study) Hippocampal firing patterns replay patterns present during the day Cueing sleep-dependent consolidation When odor was wafted during learning AND during sleep, subsequent learning is 10% better Role of PFC PFC = helps determine what we store and what we don't store Attention, judgement, cognitive control Directed forgetting = trying to forget certain associations Retrieval Cue-dependent nature of retrieval = Probability that I'l remember a past experience is heavily dependent upon the similarity of the cues that were present during encoding and cues present during retrieval  State-dependent memory = internal states (like mood or being drunk) should also match for encoding/retrieval Current mood/emotional state can bias what we remember Pattern completion and recapitulation = replaying perceptual features of what happened during an event back to the cortex Occurs when we're consciously remembering things Happens during sleep The same structures that were activated when we originally perceived something are active when we remember it Even happens when we remember what we've imagined Pair coding response 1. Find a neuron that prefers a certain abstract pattern 2. Pair that pattern with another pattern the neuron doesn't prefer 3. Pair coding response = the neuron will end up preferring the paired pattern as well as the original one after learning 4. The MTL is responsible for these stored associations If you lesion the MTL, the stored association goes away, but the orginal response remains the same Role of MTL in cortical recapitulation Non-preferred stimulus has a significant lag because it has to go through more neural processes It's happening in response to the preferred stimulus Have to first bring online representations of the preferred stimulus It's not exactly retrieving the representation of the un-preferred stimulus, but it's reinstating what was around during learning Recollection vs. familiarity Unitary view = there's a single source of information for judging is something is familiar or recollected Active signal detection = interpreting a signal to decide if you something's just familiar or if you recollect it There's a threshold, above which it's recollection and below which it's familiarity False alarm = false familiarity Dual process model = there are two different kinds of memory signals: recollection and familiarity Recollection = associative/conjunctive information Supported by hippocampus Requires cognitive control Attention at time of encoding is important Familiarity = use recollective information to decide if something is old or new MTL cortex is responsible Undifferentiated Becoming famous overnight study When names were familiar to participants because they'd heard them before, they often made the incorrect judgement that the names were of famous people This was most pronounced when Ps heard the names under divided attention, because they won't be fully attended, so things will be familiar, but not recolected Remember/know test Remember = recollection of details that surrounded the object that was studied Hippocampus is only active during correct remembers Know = don't remember the details, but have a strong sense of familiarity Hippocampus doesn't differentiate between false alarms, misses, or correct knows Self-reports as a proxy for determining whether items are being remembered with recollection or familiarity Repetition suppression and item familiarity There's a decrease in firing rate of the MTL when it becomes more familiar That's how you detect if something is familiar or not False memories Recovered memory syndrome = an individual encoded and consolidated a memory, but wasn't aware of it. Then at some point, given the right cues, the person remembers. Implanting false memories Lost in a shopping mall Experimenters generated details about a false event Ps are prompted to fill in missing details Most initially didn't remember, but then came to remember Some clung to false memory even after they'd been debriefed Spilling punch on bride's dress Confidence of false memories Deese-Roediger/McDermott Paradigm and recall/recognition Critical lure = stimuli that wren't actually studied, but are semantically related High probability of recognizing and generating these items Studied associates = what they actually studied Very high confidence on false alarms Remember and know studied items and critical lures about the same amount Altering actual memories Misinformation effect = competition between two memory traces: actual memory and misinformation Sometimes we're led astray and endorse false information People can have difficulty remembering where they acquired knowledge In eyewitness trials Memory is repeatedly probed, and each act of retrieving is a learning event Looking at slides Following watching slides, memories are probed Neutral probes = question that don't insert false information Misleading info = question gets some facts wrong; question itself brings up false information Suggestion effect = competition between what we actually learned and what suggested cues are Video of a car accident Use different language to describe the collision Influenced how fast they thought the cars were going and the details they recalled from the accident Imagination as false memory Imagination inflation = Imagining things makes people more confident that something actually occurred Rich vivid imagery creates a memory that closely resembles the contents of a real memory During retrieval, the reactivation of details gets falsely attributed to having actually seen something Source memory errors = false memories due to reality monitoring/internal-external source theory External source discrimination = trying to decide where you got a memory Internal source discrimination = trying to decide if you just thought a memory or if you actually heard it somewhere Cryptomnesia = a person mistakenly thinks that his current thoughts are novel or original Reality monitoring Role of frontal lobes Important for figuring out where information has been encountered Frontal lobe damage --> source memory errors Source memory errors Similarity-based memory errors = false remembrance based on similarity (overlap) between current stimulus and past experiences Experimental evidence People are most likely to think they recognize things when they're similar to a prototype False memory and gist memory Neural bases of true and false memory Distinguishing true from false Hippocampal response is indistinguishable for true and false memories MTL cortex isn't fooled by false memories False recapitulation = you see the same sorts of brain signals when you're remembering something as when you actually experienced it Forgetting = instances in which people have had an experience, have encoded the experience, and in the future are unable to remember that prior experience that had been encoded Decay = the weakening of memory traces due to the passing of time Ebbinghaus's forgetting curve Forgetting is negatively accelerated File drawer theory = LTM is capacity-constrained, and the more stuff you try to file, the more you lose Cue-availability = forgetting occurs due to the loss of available retrieval cues Ratio rule and global matching models Probability of remembering a certain item in a certain context depends on the strength of the item you're trying to remember and the cues in the particular context and the strength of the context and ALL the cues it's associated with Having additional cues can be really helpful because it can narrow the competition from other cues Interference = Disrupted target memory retrieval due to the presence of competing memories Retroactive interference = negative effects of later learning on memory for earlier associations Example: hard to remember old password Proactive interference = negative effects of earlier learning on memory for later associations Example: hard to remember new password Cue-overload = as more and more cues become associated with a certain cue, they become less and less effective at cueing certain memories Theories of interference Occlustion/blocking = we forget because other memories block access to larger memory The memory is just blocked, but it's still there Depends upon relative strengths of associations between that cue and other memories Output interference = the act of retrieval decreases the likelihood of being able to remember other associates of the cue Suppression = the retrieval of a competitor results in the direct suppression of another memory such that it is less accessible for recovery Retrieval-induced suppression = there is selective attention when we're trying to remember something that's associated with a certain cue during retrieval. Experimental evidence 1. Study phase- form a bunch of cue-target associations 2. Selective retrieval practice- only retrieve about half of what you learned 3. Test phase Retrieval practice items Competitor items = selected against because they didn't get retrieval practice Baseline = weren't seen in study or practice Results: accuracy practiced > baseline > competitor Strong/weak manipulation Retrieval-induced forgetting only occurs for items that serve as strong competitors during practice E.g. banana is a stronger competitor than guava Cue-independent impairment = impairments should generalize to other cues other than the ones used to guide retrieval practice Selective retrieval from memory Each act of retrieval suppresses other representations Mere act of retrieving results in a lower probability of remembering other aspects How to reduce this effect Insert a delay Think of differences Role of cognitive control and the PFC in suppression ACC = detects conflict between competing responses and signals a need for greater DLPFC control DLPFC = implements top-down modulation Think/no-think paradigm The more often Ss tried to stop retrieval, the worse recall became later Cue-independent Used for response-override mechanisms- overriding predominant memories Magnitude of forgetting is related to a decline in PFC cognitive control demands during repeated target memory retrieval Seven Sins of Memory: STAB BPM Suggestibility = memories that are implanted as a result of leading questions or comments during attempts to recall past experiences Source misattributions False and recovered memories False confessions Transience = decreasing availability of information over time Forgetting curve Adaptive role: brain doesn't hold onto useless information; optimally estimate the odds that a memory trace will be needed; adapting to environment Absent-mindedness = inattentive or shallow processing that contributes to weak encoding Depth-of-processing effect Change blindness Deep processing requires PFC Failures of prospective memory = when people forget to carry out a particular task or function Event-based tasks = remembering to perform a future action when a specified event occurs Time-based tasks = remembering to perform an action at a specified time Adaptive role: prevents having an overwhelming clutter of information Blocking = temporary inaccessibility of information that is stored in memory Tip of the tongue state = people are unable to produce a word or name, but they have a powerful subjective conviction that the items are available in memory Ugly sisters = incorrect items that are related to the sought-after target and that recur intrusively during the retrieval attempt Suppression Adaptive role: block out incessant coming to mind of numerous competing traces Bias = retrospective distortions and unconscious influences that are related to current knowledge, beliefs, and expectations Consistency bias = people's recollections tend to exaggerate the consistency between their past and current attitudes, beliefs, feelings Rating issues 10 years apart If views changed, they believe that their past views were closer to their current ones Retrospective bias = when people believe they've changed over time, they may be biased to overestimate the differences between current and past attitudes If nothing actually has changed but people think it has, they misremember their previous attitudes/skills Racial biases Adaptive role: schemas help us organize things; enhance appraisals of current self Persistence = information or events that we can't forget, even though we wish we could Patients with PTSD have a loss of cognitive control with respect to forgetting Telling people not to think about something can create a rebound effect Memories are influenced by current mood and emotions Adaptive role: traumatizing events are important to remember for survival Misattribution = attributing a recollection or idea to the wrong source Source confusion Internal External False recall Deese-Roediger/McDermott paradigm False fame effect Distinctiveness heuristic = expectation that recognition of studied items should be accompanied by recollection of distinctive details If people think about the details of things, rather than the gist, they're more likely to remember what they actually encountered Adaptive role: memory gist is important for categorization and comprehension