This preview shows page 1. Sign up to view the full content.
Unformatted text preview: Remembering and Forgetting Cognitive Neuroscience Learning and Memory "Learning is the process of acquiring new information, while memory refers to the persistence of learning in a state that can be revealed at a later time" (Squire) The study of how the brain remembers is largely a study of how the brain learns memory systems reflect learning systems. Distinctions in memory n Time scale of storage: q q q Sensory memory Shortterm memory (STM) working memory Longterm memory (LTM) Procedural/declarative Implicit/explicit Encoding Storageconsolidation retrieval n Types of memory (knowledge): q q n Stages of memory q q q TimeScale: Sensory Memory n Information available for brief periods, over ridden by incoming info Visual: q Auditory: q iconic (500 ms) echoic (several seconds) n Evidence: full/partial report Full/Partial Report Full report: Ss reports 7 +/ 2 Partial report: Ss reports 4 Regardless of line probed (doesn't know ahead of time which line will be probed) Short term/long term memory: KF/EP KF (Shallice & Warrington,1970) 28 years old n Left hemisphere damage to cortical areas: parieto occipital, following a motorbike accident at age 17 n Normal IQ n Digit span=1 n Normal memory ability otherwise n n n Can remember events, info from before his accident Can learn and remember events, info after his accident Patient E.P. Amnesic since 1992 (72 years old) n Retired lab technician n Herpes simplex encephalitis n Severe damage to medial temporal lobes n Little nonMTL damage n IQ: 103 n Digitspan 6.6 (vs controls: 58), n Spatial span 5.5 (vs controls: 47)
n E.P. CONTROL Stefanacci et al. (2000) J Neurosci Direct copy vs. (1015 min) delayed recall Mildly amnesic LTM/STM EP and KF provide a double dissociation of STM/LTM n What does it imply about the relationship between STM and LTM? n Hypothesis 1: q Information is first held in STM and then transferred to LTM STM and LTM are relatively independent learning and memory systems n Hypothesis 2: q Evidence for STM/LTM distinction n Since LTM memory can be normal with abnormal STM memory (e.g., patient KF), information does not have to go through STM to get to LTM. Rehearsal Maintenance of STM memory requires rehearsal. q Entry into LTM does not require rehearsal.
q n STMWorking memory n STM is a different system Working Memory (WM): q q The process of maintaining an representation of information so that it is available for use. After the information is used, it is typically discarded. Baddeley Three Components of Working Memory n n n Phonological Loop (Left temporoparietal digit span 7 2 items) Visuospatial Sketchpad (bilateral occipitoparietal 34 objects ) Central Executive (bilateral frontal lobes) WM is based on activation of represented items. LTM is based on synaptic changes. q Rehearsal effects suggest that n n Distinctions in memory n Time scale of storage: q q q Sensory memory Shortterm memory (STM) working memory Longterm memory (LTM) Procedural/declarative Implicit/explicit Encoding Storageconsolidation retrieval n Types of memory (knowledge): q q n Stages of memory q q q HM (Scoville & Milner,1957,etc.) n n n Bilateral medial temporal lobectomy (hippocampi) Age 29 to alleviate chronic, severe epilepsy Normal IQ Normal digit span Like EP: can't form new, lasting memories (can't remember a task he's performed before and if the experimenter leaves the room can't remember they've ever met) n n Stefanacci et al. (2000) J Neurosci H.M. Scoville & Milner (1957) n n H.M.'s recall of old memories is relatively intact Therefore, his memory performance represents a distinction between types of learning or formation of new memories q q + new STMs new LTMs Anterograde: Retrograde: inability to create new longterm memories inability retrieve old memories n anterograde/ retrograde amnesia q q n HM has anterograde amnesia What kind of lasting memories can't be created? Procedural/Declarative H.M. not all types of new, longterm memory creation were affected q Can learn motor and cognitive skills (but has no knowledge of having learned them).
q Procedural Memory H.M. q n cognitive procedures: Tower of Hanoi Move the entire stack to another peg, obeying the following rules: q q Each move consists of moving the upper disk from one of the pegs onto another peg (on top of the other disks that may already be present on that peg). A disk may only be placed on top of a larger disk. Procedural vs. Declarative n n Procedural: processes (motor, perceptual or cognitive) Declarative: facts and events (knowledge "that") n Episodic: Semantic: n knowledge of specific personal experiences, their content, time place (episodes) world, word knowledge n HM had difficulty acquiring both episodic and semantic knowledge but it has been argued that these memory types do dissociate in other amnesic individuals. q q Selective difficulties with episodic knowledge Selective difficulties with semantic knowledge Types of Memory or Memory Retrieval: Implicit/Explicit knowledge that can be retrieved without conscious recollection n Explicit: knowledge that can be retrieved only with conscious recollection n Neurologically intact individuals (as well as amnesics) show many dissociations in terms of implicit/explicit memory abilities
n Implicit: Normal subjects: Implicit/Explicit Explicit oral report measure: Subjects report digits were presented in completely random sequences Implicit RT measure: RTs to repeating sequences become faster as compared to random sequences Priming (implicit) vs. Recognition (explicit) tests n n n E.P. studied a list of 24 words (e.g,. WINDOW was one of the items) After a five minute delay, he was given priming (implicit) and recognition (explicit) memory tests Implicit test of memory for studied words: q Word stem completion: Shown the word stem "WIN__" asked for the first word that came to mind. q Just like healthy controls, he often completed the stem with the studied word "window". Does this show memory for the studied word? n His priming score (25%) compares the likelihood of completing the stem to a studied word vs. likelihood of control Ss that did NOT study the words n Explicit test of memory for studied words: q Recognition memory: Immediately after he completed the stem, he was shown him two words WINDOW and WINTER and asked which one he'd seen a few minutes earlier. q Despite having just said "window" in stem completion he scored at chance . n Repetition Priming .53 The memory responsible for repetition priming on many trials, j allowed EP to generate the correct word. Yet, that information was totally unavailable to him when making his recognition memory judgment Recognition .40 100 80 % Correct .30 Priming 60 40 20 0 .20 .10 0 Controls EP Controls EP Stark & Squire (2000), Behav Neurosci Explicit/Implicit Declarative/Procedural n Explicit = Declarative q Conscious access to items (facts, words, events) n Implicit Procedural cognitive/motor skills q Priming activation of items, not necessarily leading to conscious access
q Memory Transient Sensory Visual Echoic Haptic Longterm Working Articulatory loop Visuo Habits Perceptual Classical spatial sketch pad Semantic/episodic Learning Conditioning Declarative Implicit (Explicit) Skills & Priming & Simple Nonassociative Learning Facts & Events Central Executive Adapted from Squire & Zola Distinctions in memory n Time scale of storage: q q q Sensory memory Shortterm memory (STM) working memory Longterm memory (LTM) Procedural/declarative Implicit/explicit Encoding Storageconsolidation Retrieval n Types of memory (knowledge): q q n Stages of memory q q q Memory stages Encoding n Consolidation/storage n Retrieval n n HM--difficulties specifically in memory consolidation q Although his amnesia was primarily anterograde, there was a retrograde component n Memory of events prior to surgery was WORSE for knowledge acquired closer (rather than further) in time relative to the surgery Consolidation n Retrograde amnesia: Focal retrograde amnesia: the primary deficit is a loss of remote memory q Graded retrograde amnesia: memory loss inversely related to time of learning implies consolidation period
q Graded retrograde amnesia: Memory loss inversely related to time of learning n n n n Train an animal on a task Perform hippocampal or control lesion X days after training (After recovery from surgery) test animal on the task Figure depicts accuracy at test Days between training and surgery Squares: control Circles: Hippocampal Consolidation n n n Retrograde memory loss inversely related to time of learning Implies consolidation Implies the medial temporal lobe structures are critical for consolidation Memory stages Encoding n Consolidation/storage n Retrieval
n n n n n n n n n n n n n n n n table sit legs seat couch desk recliner sofa wood cushion swivel stool sitting rocking bench n n n n n n n n n n n n n n n hill valley climb summit top molehill peak plain glacier goat bike climber range steep ski n n n n n n n n n n n n n n n bed rest awake tired dream wake snooze blanket doze slumber snore nap peace yawn drowsy Encoding n Experience to be encoded: q Went with friend X to movie Y at theatre Z on day W n Consists of many features distributed across different neural systems q q q Sensory (sights, sounds, smells) Conceptual (thoughts) Emotional Feature binding: for a complete memory Pattern separation: to avoid confusion with similar memories n Encoding involves: q q Retrieval 1. Retrieval cue: q movie Y Retrieval of features associated with movie Y (friend X, theatre Z, etc.) Is it Veridical? Incomplete? Fantasy? Degree of familiarity, vividness, detail 2. Pattern completion: q 3. Evaluation of the pattern: q 4. Criteria for evaluation q RECOGNIZE THE WORDS couch molehill flower chair awake dinner plate peace table picture valley bed drowsy mountain car sleep table sit legs seat couch desk recliner sofa wood cushion swivel stool sitting rocking bench bed rest awake tired dream wake snooze blanket doze slumber snore nap peace yawn drowsy hill valley climb summit top molehill peak plain glacier goat bike climber range steep ski False memories n Failure in encoding: q Insufficient memory separation Low criteria for accepting a memory as veridical n Failure in retrieval: q Remembering and forgetting n "Remembering is not simply a process of passively activating stored information. Instead, people's expectations, metamemorial beliefs, even how they are instructed to examine their memory, influence the kind and amount of information that is remembered" (Dodson & Schacter, 2001) Memory as a dynamic system "In a 1992 oped piece for the New York Times, Garry Trudeau recollected some of his experiences related to the draft for the Vietnam War. Trudeau remembered receiving calls of concern from friends and family on the night of the draft lottery, after they had heard about his low number. He then recalled a series of events involving his attempts to gain a draft deferment: requesting a national security deferment from the draft board for his work with a magazine deciding not to apply for conscientious objector status because he could imagine circumstances in which he would take another's life preparing for his interview with the draft board by receiving a "memorable haircut" and finally, gaining a medical deferment from the board after sending them, upon his physician father's advice, Xrays revealing a past ulcer. This recollection, he reported, remained unchanged for 20 years. However, after talking to others and examining the records of his draft correspondence, Trudeau uncovered some notable discrepancies between his recollection and what actually happened. No family member or friend remembers making a call of concern. Trudeau now believes he imagined their concerns, because the act of examining this recollection led him to remember that he was in fact out having a few beers that night. He discovered that he actually applied for an occupational deferment and, upon reflection, wonders how he could have believed that working for a "glorified travel magazine" was justification for a national security deferment. He also never received a "memorable haircut", not did he apply for conscientious objector status in part because of the prohibitive paperwork" (Dodson & Schacter, 2001) Memory associated with (nonfrontal) neo cortical damage Focal retrograde amnesias n Memory loss in specific domains n q Language, object recognition, navigation, number processing, etc. False Memories "To what extent is the life we remember the knowledge and expectations we have, and the self we seem to ourselves to be, a product of experience and to what extent a product of our imagination?" (Johnson, 1985) Distinctions in memory n Time scale of storage: q q q Sensory memory Shortterm memory (STM) working memory Longterm memory (LTM) Procedural/declarative Implicit/explicit Encoding Storageconsolidation retrieval n Types of memory (knowledge): q q n Stages of memory q q q Neural Bases of learning and memory n n n Medial temporal lobes Frontal lobes Neocortex Medial temporal lobe: limbic structures Amnesias resulting from MTL damage Evidence from HM and others n n n n n Cognition and memory are separate. The ability to acquire new memories is a distinct cerebral function located in the medial portions of the temporal lobe. The medial temporal lobe is not required for immediate (working) memory. The medial temporal lobe cannot be the ultimate storage site for longterm memory. The ultimate site/s for longterm memory includes neocortex Amnesia resulting from frontal lobe damage n n Associated with high false recognition rates (false memories) Patient BG (Schacter, et al., 1996) q False recognition of new/studied words 50% (controls 17%) n Ongoing research to understand to what extent this is due to encoding or retrieval failures (or both) n n If encoding failure: expect improvement with instructions and strategies for more efficient encoding If retrieval failure: expect improvement with instructions for more stringent retrieval criteria Frontal lobes and executive function n Executive functions: Working memory (lateral prefrontal cortex) q Planning q Monitoring progress to goals q Response selection and inhibition q Task switching q Encoding q Retrieval
q Encoding and retrieval in the frontal lobes The relationship between activity in the frontal lobes and memory strength Wagner et al. 1998 n n n n fMRI scanning at the time of encoding (study of words) Subjects were then tested for their memory of the words outside of the scanner Sorted fMRI data from the encoding period according to whether or not the word was correctly or incorrectly remembered later Compared fMRI encoding data for correctly vs. incorrectly remembered words n Brain areas that show differences in activation at time of encoding, predict accuracy at time of test Evoking memories in lateral temporal cortex Moriarity, et al., (2005) n n n 34 yearold man with intractable epilepsy underwent lobectomy that removed left anterior temporal regions and amygdalahippocampus Seizures returned cortical grid implanted to identify seizure foci Cortical stimulation Neural Bases of learning and memory n MTL and diencephalon (thalamus & mamillary bodies): n n Creating and consolidating declarative memories Explicit memory retrieval creating new procedural memories Implicit memory retrieval Encoding and Retrieval n Neocortex: n n n Frontal lobes: n Stability vs. Plasticity Dilemma in Declarative Memory n Problem: q n Solution: q q q q Need to learn new things quickly New learning should not disrupt previously learned things. Modeling using distributed representations shows that rapidly learning new things disrupts representations of previously learned things. q Two step learning: First learn quickly, using local representations (encoding in hippocampus) Then use these local representations to slowly change distributed representations in a way that does not disrupt existing knowledge (consolidation in cortex) McClelland, McNaughton & O'Reilly (1995) Two learning systems: neocortical and hippocampal n Hippocampus: q q q learns quickly rapid, onetrial learning local representations during consolidation serves as "teacher" to the neocortex, until neocortical connection changes are robust n Neocortex: q q q learns slowly, by making small adjustments to connections among neurons allows integration into complex knowledge structures distributed representations n Simulations show that this combination solves the stability vs. plasticity dilemma Neural Basis of Procedural Learning n Striatum q Caudate Nucleus + Putamen n Frontostriatal circuit important for sequence learning Learning q q n Repetition Trial and Error, Reinforcement Procedural vs Declarative Learning in Language n Past Tense of Verbs q Regular: + ed n n Rule/ Procedural knowledge Basal Ganglia ? Fact, Declarative knowledge Neocortex ? q Irregular n n Ullman et al. (1997) n Verb generation q q q n Results for Accuracy q Irregular Regular Novel (Pseudowords) Posterior aphasia Semantic dementia (Alzheimer's) Anterior aphasia Parkinson's (underactivation of basal ganglia) Huntington's (overactivation of basal ganglia) Irregular < Regular, Novel n n Five patient groups q q Posterior Aphasia, Dementia Anterior Aphasia, Parkinson's, Huntington's q q Regular, Novel < Irregular n q q Only Aphasics, Huntington's made over regularization errors (e.g., digged, lookeded) q Memory Transient Sensory Visual Echoic Haptic Longterm Working Articulatory loop Visuo spatial sketch pad Facts & Events Declarative Implicit (Explicit) Skills & Priming & Simple Nonassociative Habits Perceptual Classical Learning Conditioning Learning Central Executive Emotional Skeletal Responses Responses Association Areas Prefrontal cortex Basal Ganglia Sensory areas in neocortex Amygdala Cerebellum Reflex Pathways Neocortex Initia l l y Medial Temporal Lobe & Diencephelon ntua Ev e lly Neocortex ...
View Full Document
- Spring '08