Working Memory I

Historical Foundations and Terminology of Memory Research

• Primary versus Secondary Memory (William James, 1842–1910)

  • William James introduced a foundational distinction between primary memory and secondary memory based on introspection and the systematic analysis of his own conscious experiences.

  • Primary Memory: Refers to memory for "just-past" events. Information in this store remains continuously in conscious awareness without being stored away in long-term memory.

  • Example: When listening to the end of a word, the beginning of that word is currently held "in memory" rather than retrieved from storage.

  • Secondary Memory: Defined as "memory proper," referring to knowledge or past experiences that have left active consciousness and are subsequently recalled back into awareness.

• The Behaviorist Era

  • In the early $1900\text{s}$, the behaviorist movement explicitly rejected introspection and argued against the scientific investigation of conscious experience.

  • Consequently, scientific interest in primary memory and short-term memory research declined significantly during this era.

• The Cognitive Revolution and George A. Miller (1956)

  • George A. Miller challenged behaviorist doctrine by investigating how individuals internally represent and process information.

  • Miller conceptualized humans as information-processing systems with a specific "channel capacity."

  • Wi-Fi Analogy: Transmitting a high-resolution photo across weak Wi-Fi limits the total quantity of data successfully delivered.

  • Miller famously described being "persecuted by an integer," specifically referring to the value $7 \pm 2$.

  • 
    

  

  • Miller's theoretical framework rested on two primary experimental lines:

  1. Absolute Judgment Experiments:

     • Experimental Setup: Subjects were presented with isolated sensory stimuli (e.g., specific auditory pitches, tastes, or visual lights) and instructed to assign a categorical label identifying their intensity (e.g., loudness, saltiness, or brightness).

     • Findings: Categorization is simple when using $2$ categories, but performance drops rapidly once reaching approximately $7$ categories.

     • Quantitative Capacity: The channel capacity for absolute judgment is limited to approximately $7$ categories, corresponding to $2.5\text{ bits}$ of information.

     • 
       

     

     (Tested across pitch frequencies ranging from $100\text{--}8000\text{ CPS}$).

  2. Immediate Memory Span Experiments:

     • Experimental Setup: Participants were presented with sequences containing $X$ items (e.g., numbers, letter strings, or full words) and asked to recite them back shortly after.

     • Information Content vs. Capacity: Words carry significantly more raw information than binary digits (e.g., a word sequence like dog sheep cat cow deer horse versus a binary sequence like 0 1 0 1 1 0), yet individuals retain approximately $7$ items in immediate memory regardless of item complexity.

     • Core Conclusion: The span of immediate memory is limited by chunks (meaningful groups of information that can be decomposed) rather than by raw bits of information. Working memory capacity is fundamentally defined by chunks.

• Current Multi-Store Models of Memory

  • The Atkinson-Shiffrin ($1968$) multi-store model serves as the foundational structural framework for memory processes.

  • 
    

  

  • Structural Component Architecture:

  • Sensory Input: External information enters Sensory Memory. Unattended sensory information is rapidly lost.

  • Attention: Selective attention transfers information into Short-Term Memory. Unrehearsed information in short-term memory is lost.

  • Maintenance Rehearsal: Mental repetition maintains items within short-term memory.

  • Encoding & Retrieval: Encoding transfers items into Long-Term Memory, which can be accessed back into short-term memory via retrieval (some long-term information may be lost over time).

  • Working Memory Definition: Refers to the short-term storage store combined with rehearsal and active processing, emphasizing its dynamic operational nature.

Sensory Memory Dynamics and Paradigms

• Definition and Overview

  • Sensory memory functions as an extremely brief buffer where raw sensory input is stored for a fraction of a second before decaying.

• Sperling's (1960) Visual Memory Paradigm

  • Sperling evaluated the capacity and duration of visual sensory memory by briefly presenting visual arrays of letters for a fraction of a second.

  • Whole Report Method:

  • Participants were asked to recall all letters from a $12$-letter display.

  • Result: Participants recalled an average of $5/12$ items ($42\%$ accuracy).

  • Partial Report Method:

  • Participants were cued immediately after presentation to recall only a single random row ($4$ letters).

  • Result: Participants consistently recalled $3/4$ letters ($75\%$ accuracy) from the cued row.

  • Theoretical Implications:

  • Individuals retain substantially more information in sensory memory than whole-report measures capture.

  • Visual information decays so rapidly that the trace fades before participants can report the full display during whole-report tasks.

  • Subtypes of Sensory Memory:

  • Iconic Memory: Brief sensory memory for visual information.

  • Echoic Memory: Brief sensory memory for auditory information.

• Sensory Memory Visual Display Demonstrations

  • Display Matrix 1:

  • H T F R

  • I >

  • V Q E F

  • O N

  • D Z N Y

  • Display Matrix 2:

  • A E B F

  • W T H M

  • X G S L

  • Display Matrix 3:

  • P R Y S

  • D H I J

  • Q T L X

  • Display Matrix 4:

  • E F G J

  • X I P L

  • M S U R

  • Display Matrix 5:

  • I T G D

  • Q E M L

  • R K O W

  • Display Matrix 6:

  • O P A S

  • F T Y

  • Y U

  • G B E C

Working Memory Capacity and Limits

• Characteristics of Working Memory

  • Working memory is a limited-capacity system that maintains information consciously accessible for short periods of time.

  • Comparison with Long-Term Memory:

  • Long-term memory holds a virtually unlimited amount of information.

  • Long-term memory operates without requiring continuous conscious awareness.

  • Long-term memory stores information across indefinite periods of time.

• Measuring Working Memory Capacity

  • The Magical Number Seven: Miller's ($1956$) estimate of $7 \pm 2$ chunks serves as a benchmark for working memory capacity.

  • Digit Span Task:

  • Measures capacity by requiring immediate recall of random digit sequences.

  • Memory task difficulty increases significantly once sequences exceed approximately $7$ digits.

  • Sample Sequence Progression:

    • $5$ digits: 9 0 3 6 7

    • $6$ digits: 4 7 6 2 9 1

    • $7$ digits: 3 7 8 5 2 6 0

    • $8$ digits: 1 9 8 7 2 0 4 6

    • $10$ digits: 1 2 0 7 3 9 6 8 2 9

  • Historical Note: Standard telephone numbers were originally formatted with $7$ digits to match human digit span capacity limits.

  • Corsi Block-Tapping Task:

  • Measures visuospatial working memory capacity.

  • Demonstrates that spatial memory exhibits identical limits: participants experience difficulty recalling sequences of more than approximately $7$ highlighted blocks.

  • 
    

  

• Factors Moderating Capacity Limits: The Word Length Effect

  • Capacity is influenced by the physical and temporal characteristics of information items.

  • Word Length Effect: Working memory holds more chunks when the words take less time to pronounce during mental rehearsal.

  • Short Word Sequence: Prague, Nice, Rome (fewer syllables and shorter pronunciation time allow higher span retention).

  • Long Word Sequence: Bratislava, Barcelona, Manchester (polysyllabic words with longer pronunciation durations decrease effective memory span).