Prof. Greg Francis 5/23/081Working memoryIIE 269: Cognitive PsychologyGreg FrancisLecture 17A problem with IQ tests.Working memory Current thought, awareness extension of short-term memory small capacity rapid forgetting Processor of information not a storage device hypothesizes mechanisms that lead to memorypropertiesCentralexecutiveVisuo-spatialsketchpadPhonologicalloopPhonological loop Two components Articulatory control process(ACP)» converts non-speechinformation intospeech code» rehearsal / refresh Phonological store (PS)» similar to how we firstdescribed STM (itemsdecay from memory)ACPPSPhonological loopACPPSPhonological loopLoop capacity How many items can bekept in the phonologicalloop? Depends on two factors Duration before decay fromPS Speed of rehearsal Spinning coins!Loop capacity Measure memory span (s) Measure verbal rehearsalrate (r) Estimate duration of decayin PS (t) t = s/r Get an estimate of t=1.6secondsACPPSPhonological loopLoop capacity Memory in the phonological loopis a combination of how longthings stay in the PS and howfast the ACP can rehearse Consider a list of four letters visual presentationACPPSPhonological loopBBB BBBBProf. Greg Francis 5/23/082Loop capacity Memory in the phonological loopis a combination of how longthings stay in the PS and howfast the ACP can rehearse Consider a list of four letters visual presentationACPPSPhonological loopB FFF FFFFBBBBBLoop capacity Memory in the phonological loopis a combination of how longthings stay in the PS and howfast the ACP can rehearse Consider a list of four letters visual presentationACPPSPhonological loopB F HF FFFFBBBBBHHHHHHFLoop capacity Memory in the phonological loopis a combination of how longthings stay in the PS and howfast the ACP can rehearse Consider a list of four letters visual presentation if B is too weak to initiate rehearsal,it is forgotten!ACPPSPhonological loopB F H NF FFFFBHHHHHFNNNNNN HPhonological similarity Memory of a list of items is worsewhen the items sound the sameBetterrecallWorserecallB F H NB G P TPhonological similarity All items are stored in phonological loop similar sounding items interfere with each other in the PS two possibilities:» 1) harder to rehearse (effect in the ACP)» 2) fade more quickly (effect in the PS)ACPPSPhonological loopBGP TACPPSPhonological loopBFHNArticulatory suppression Subjects sees (hears) a list of phonemes Also repeats a phrase over and over e.g., “tippy-toe, tippy-toe, tippy-toe,... Recall is worse Much bigger effect for visual presentation than for auditorypresentation for auditory presentation, one still finds a phonologicalsimilarity effectProf. Greg Francis 5/23/083Articulatory suppression Repeating phrase ties up the ACP Without rehearsal more forgetting occurs Also limits transfer of visual information into phonologicalinformation (reading)ACPPSPhonological loopBFHNVisualtippy-toeAuditoryArticulatory suppression Repeating phrase ties up the ACP Studies still find a phonological similarityeffect We suggested two possibilities:» 1) harder to rehearse (effect in theACP)» 2) fade more quickly (effect in the PS) Since the phonological similarityeffect is there even when the ACP isnot involved, it must be possibility 2(in the PS)ACPPSPhonological loopBGP Ttippy-toeEffect of rehearsal rate Capacity of thephonological loop dependson the rate of rehearsal (r) A set of items that takeslonger to rehearse is harderto get through more likely that some itemswill drop out before you getback to the first itemACPPSPhonological loopBFHNWord length effect Memory span is related to the length ofwords4.17 2.8CHADBURMAGREECECUBACZECHOSLOVAKIASOMALILANDNICARAGUAAFGHANISTANWord length effect The rehearsal properties of the phonological loop suggest thatwhat matters is the time required to say the word, not itswritten length There is some evidence that this is true (but controversial) E.g, if experiment measures percent correct recall89%79%DECORHACKLEWIGGLEPEWTERZYGOTEVOODOOCOERCEHUMANECoglab data The CogLab experiment on memory span showsdata in agreement with our expectations (145subjects) Item type Final list length Numbers 6.6689 Letters that sound different 6.0068 Letters that sound similar 5.1103 Short words 5.5379 Long words 4.4 If it is just the rate of rehearsal, we should be able tomake some other predictionsAbout 7Phonologicalsimilarity effectWord lengtheffectProf. Greg Francis 5/23/084Language effects Somelanguages arespoken morequickly thanothers Should allowlarger memoryspan it doesIQ’s Ellis & Henley (1980) investigated complaints about WISCintelligence scores Welsch children tended to score lower thanEnglish children Part of exam checks memory span and the slower rate of speech in Welschpartly explains the difference bilingual Welsch students tested in Englishgot better scores than when tested in WelschIrrelevant speech effect Does irrelevant “background” sound affectmemory? E.g., studying with the TV on Three groups of subjects recall consonants 1) no background 2) background = nonsense words 3) background = noise burstsbestworstIrrelevant speech effect The presence of phonemes in thebackground is critical to the effect strong effect when background is spoken inGerman, even for English speakers Suggests that background phonemesinterfere in the PS Study with classical music if you needsomething!Reading span Daneman & Carpenter (1980) subject must recall last word of each sentenceStudents in PSY 200 love to take exams.Autumn exploded in a brilliant array of reds and greens.The car skidded across the pavement and slammedinto the utility pole.Reading span See how many sentences they can workwithS1S1, S2S1, S2, S3S1, S2 , S3, S4S1, S2 , S3, S4, S5Prof. Greg Francis 5/23/085Reading span A high reading span (close to 7) indicates agood ability to keep things in workingmemory w/o getting confused They are also good readers they understand what they read betterReading span Second part ofstudy Subjects reada paragraphand answerquestionsabout middlesentences higher readingspan = betterunderstanding!Working memory theory Good unified account of a number of interesting effects applied to lots of situations involving