pl3103-week4-stm-handout.pdf

PL3103 Cognitive Psychology Dr. Cynthia Siew Today’s Topic: Short-term-memory (Week 4) After today’s lecture, you will learn about: 1. The similarities and differences between short-term mem- ory and working memory. 2. The methods used to measure STM and WM capacity....

PL3103 Cognitive Psychology Dr. Cynthia Siew Today’s Topic: Short-term-memory (Week 4) After today’s lecture, you will learn about: 1. The similarities and differences between short-term mem- ory and working memory. 2. The methods used to measure STM and WM capacity. 3. The key theoretical models that explain how STM and WM work. Atkinson & Shiffrin’s model of memory (1968) There are 3 types of memory: 1. Sensory memory 2. Short-term memory 3. Long-term memory 1 Sensory memory stores information that we receive from the environment – visual information in an iconic store – auditory information in an echoic store has large capacity (full perceptual detail is preserved) but decays rapidly Short-term memory only limited information from sensory memory is trans- ferred to STM – c.f. Broadbent’s model of attention from Lecture 3/FoC Ch 3 – information not attended to will be lost STM has limited capacity Long-term memory rehearsal is necessary to transfer information in STM to LTM – information in STM easily displaced by new incom- ing sensory information – rehearsal = “subvocal repetition of verbal material” Evaluation This is very dated model! But it is relevant for setting the theoretical framework for memory research. Nevertheless we should also be aware of its shortcomings. 2  1. Not true that unattended information is never processed into STM: We know from research in attention (Ch3/L3) that unattended stimuli can be processed without our awareness. 2. Only allows 1-way flow of information from SM -> STM -> LTM: We know from research in perception (Ch2/L2) that bottom up and top down information (i.e., knowl- edge stored in LTM) can interact to influence perceptual processing. 3. Focus on storage capacity vs. processing of stimuli 4. Coarse-grained view of LTM (unitary store, rehearsal as the only form of learning permitted): Next lecture we will learn more about the LTM and see that it is a lot more complex than Atkinson & Shiffrin suggested Measuring STM capacity How many items of information can we store? How long can the information stay? How many? Memory span is the number of items that one can recall imme- diately in the correct order of its presentation. General approach: Present items in a random (non-meaningful) order, see how many can be recalled. In these studies, the items could be letters, digits, words… 3 7+2 or 4+1? According to Miller (1956), the STM capacity is about 7 items, plus or minus 2. According to Cowan (2001), it is about 4 items, plus or minus 1. Who is right? Items or Chunks? Miller’s estimate could be an overestimation because people are able to form chunks from grouping letters/digits as a single unit, and leverage on them to increase their span. For example: AYECTETPE (9 items) vs. AYE_CTE_TPE (3 chunks) Cowan (2001)’s experimental design (with fast presentation of digits and unpredictable recall points) prevented participants from forming chunks and leveraging on their LTM. Not all chunks are the same! Example 1: Word length effect Larger word span for short words than long words. Example 2: Cultural differences in digit span Chinese speakers have a larger digit span than English speak- ers. 4 Example 3: Phonological similarity effect Memory span is smaller for words that sound like each other (key, me, ski…) than for words that do not sound like each other (cat, pot, who…). These results suggest that STM is dependent on phonologi- cal/auditory information (how quickly you can say something in your head) vs. visual information. ### Tik Tok 12345678 challenge English digits Chinese digits Primacy and recency effects The classic U-shaped serial position curve. Primacy effect: better recall for early items -> Atkinson & Shiffrin argued that this was due toe rehearsal that transferred the early items into LTM 5 Recency effect: better recall for later items -> these items were still in STM, but fragile and can be disrupted easily. disruption of late items could be due to decay (delay) or inter- ference, e.g. count backwards. From Glanzer & Cunitz (1966). How long? Memory performance diminishes rapidly even after a few sec- onds. From Peterson & Peterson (1959). Recall of 3 consonant items while What causes this rapid forgetting? counting backwards to prevent rehearsal. Decay: a natural loss of information over time Interference: from other items, previous trials, task demands 6 Evidence for both interference and decay (Ricker & Cowan, 2010) Used unfamiliar special characters (not letters) to prevent re- hearsal. Method present special character for 750 ms after 1.5s, 3s, or 6s, show a probe character participants decide if probe was shown previously or not Manipulation to induce interference: repeat digits that they heard (low load), or subtraction of digits heard (high load) Manipulation to induce decay: time between presentation of character and test (retention interval) results show evidence of both decay and interference, although most papers tend to find evidence for interference Ricker, T. J., & Cowan, N. (2010). Loss of visual working memory within seconds: The combined use of refreshable and non-refreshable features. Journal of Experimental Psychology: Learning, Memory, and Cognition, 36(6), 1355–1368. https://doi.org/10.1037/a0020356 7 Baddeley & Hitch’s Working Memory Model (1974) Going beyond storage and capacity, but also how information is processed and managed in order to achieve a goal. For example: Do the following mental sum in your head (no calculators!) 24 x 7 = ? Reflect on your internal memory processes as you are doing the mental sum. WM components Central executive: attentional controller of other components Phonological loop: storage and processing of sound informa- tion Visuo-spatial sketchpad: storage and processing of visua and spatial information Episodic buffer: storage buffer to hold information from PL, VSS, LTM 8 Phonological loop has limited storage and processing capacity for sound in- formation consists of a passive phonological store (speech percep- tion) and articulatory control process (speech production and rehearsal) important for learning new words (novel sequences of sound!) Visuo-spatial sketchpad limited storage and processing capacity for visual and spa- tial information (what and where) consists of a passive visual cache (stores visual info) and inner scribe (refresh info in cache and transfer info to CE) measured with the Corsi blocks test Central executive attentional controller of other components no storage capacity limited processing capacity for all information types 9  involves executive functions: a set of cognitive skills for control and coordination of goal-directed behaviors Executive functions Four sub-skills: 1. Task setting and shifting: activation and shifting of goals 2. Energising: initiating and sustaining a response 3. Monitoring: updating your WM representations 4. Inhibition: suppress irrelevant behavior or responses in relation to the goal Think about how each of these skills come into play in your daily life as a uni student. What are your goals and how would your EF skills help you achieve them? A recommended popular science book on this topic: On Task Episodic buffer storage buffer to hold information from PL, VSS, LTM (since the CE cannot store information) integrates different sources of information Evaluation Is it an overly complicated model? Baddeley’s model is vaguely described with little specifics on how these processes would occur Neuroscientific evidence does not find support for each of these components–brain activation is more related to the type of information involved – should there be separate components for smell and touch information? egocentric vs. allocentric spatial representations? 10 Alternative model: Unitary store model working memory is simply an activation of a section of long-term memory thanks to attentional focus there is no distinction between STM/WM and LTM Evaluation Neuropsychological evidence indicates that STM and LTM are separable memory representations (patients can have good STM with impaired LTM) Does not adequately explain how processing occurs (i.e., how information is manipulated in service of a goal) Measuring working memory capacity the task should measure more than just STM capacity includes a processing or manipulation component of the presented information Examples Reverse digit span repeat numbers backward Reading span read sentences for comprehension AND remember the last word in each sentence Operation span evaluate math equations AND remember a letter/word 11 Try out these demonstrations to measure your own working memory Try to think about what types of memory each task is mea- suring, and whether the task is focused on storage or stor- age+processing. WM Task 1 WM Task 2 WM Task 3 WM Task 4 Your data should be saved as a.csv file and downloaded to your Today’s Topic: Short-term-memory (Week 4) Downloads folder. Feel free to explore the raw data and see how After today’s lecture, you have learned about: you did :) 1. The similarities and differences between short-term mem- ory and working memory. 2. The methods used to measure STM and WM capacity. 3. The key theoretical models that explain how STM and WM work. See you next week for the lecture on “Long-term memory”! 12 13