4/20: Episodic memory: Encoding I.doc

4/20: Episodic memory: Encoding I Semantic memory = memory for facts and general information about the world Episodic memory = memory for a specific event (what) that occurred ina specific time (when) and place (where) How is episodic memory tested? Subjects study a list of words Free recall Anything you can remember Primacy and recency effects Hard: about 45% Cued recall Get particular bits of the experience by presenting a specific cue "Tell me which word on the list completes this stem: mot___" Easy: about 75% Memory is cue-dependent How well people do depends on how much support you give them Recognition Give a bunch of words, some were on the list, some weren't. Say y/n if you saw them before Stages of LTM Encoding = placing information into memory Storage/consolidation = keeping the information in a permanent store Retrieval = bringing information back to mind Encoding = incidental byproduct of the active stimulus processing carried out during an event Episodic encoding = processes that yield a durable memory trace such that an event can be subsequently consciously remembered Attention- crucial to encoding Levels of Processing (LoP) The power of retrieval for encoding Factors that maximize sort-term memory retention often result in poorer long-term retention Can lead to an overestimation of learning Optimal learning requires "desirable difficulties" What does maximize long-term retention? Retrieval during study Tests are the best way to learn thigns The generation effect = people who were asked to generate words had better memory later than those who were just asked to read the words Experiment- cued recall Read conditions: synonym and rhyme Unhappy - SAD Pad - SAD Just read them and try to remember Generate conditions: synonym and rhyme Unhappy - S__ Pad - S__ Not just read, but generate a word that fits Explanation Greater elabration/processing of meaning or phonology Greater engagement of retrieval process during study that are likely to be engaged at test Power of retrieval as an encoding event: Roediger & Karpicke (2006) Text passage to learn IVs SSSS: four study presentations SSST: hree study and 1 test No feedback on test STTT: 1 study 3 test No feedback on test Judgement of learning (how well they felt they learned it) SSSS > SSST > STTT Actual memory performance: immediate test SSSS > SSST> STTT Actual memory performance: delayed test STTT > SSST > SSSS Attention mediates the testing effect Presented with a sequence of pictures Judge whether- Living/nonliving Pleasant/unpleasant Test 1: full attention or divided attention (distracted) Later, judge whether they saw the item before or not When you saw the item, were you making living/nonliving judgement or pleasant/unpleasant judgement? Lets you see if they have a sense of familiarity or whether they remember the whole study event Test 2: 2 days later Some studied and untested Studied with full attention Studied with divided attnetion Novel Results FA > DA Spaced practice Not all encoding events are created equal Massed vs. distributed/spaced practice Spacing effect = longer lags between study presentations seem to give better memory than shorter ones Read/generate test IVs Once-presented pairs (Foot-Shoe) Twice-presented Massed Foot-Shoe Foot-Shoe Spaced Foot-Shoe -20 pairs- Foot-Shoe Results Esp. for generation, spaced presentation is a lot better Understanding distributed practice Deficient processing = during mass/short lag trials, the second occurrence of an item is not processed fully Encoding variability = The longer apart the study events are, the more variable the encoding is Rich memory trace Lots of opportunities to look for cues Estes' stimulus sampling theory (1955) Lots of different stimuli can serve as retrieval cues (external stimuli, internal body states, random thoughts, etc.) Initial state: Some are available for learning, some are unavailable for learning Probably not consciously aware of them Study 1: connect available stimuli to what you're learning This associations will help support later recall Because they're all available in the moment, probability of retrieval is 1 Lag Short Some stimuli pass out of the current available stimuli Others pass into the currently available stimuli Long More will pass in/out More different cues Study 2 Different stimuli to connect to hat you're studying Test More of the stimuli have moved into and out of the currently available stimuli After short lag: probability of remembering = .5 After long lag: probability of remembering = .75 Not all spacing events are created equal Expanding practice = Allowing the spacings to get bigger is even better than spacing them out equally Ex. 5 minutes, 15 minutes, 1 hour