CP Lecture 3 - Memory

Iconic Memory (Visual Sensory Memory)

• Capacity: Very large (essentially the entire visual field).

• Duration: Very short, decaying in less than 1 second (e.g., 250-500 milliseconds).

• Format: Pre-categorical, literal, "raw" visual information (a direct snapshot, not yet processed for meaning).

Sperling’s Experiment — Iconic Memory

Sperling briefly flashed a grid of 12 letters (e.g., 3 rows of 4).

1. Whole Report Condition: Participants were asked to recall all the letters they saw.

   • Result: They could only recall about 4-5 letters (37.5%).

2. Partial Report Condition: Immediately after the flash, a tone (high, medium, or low) told them to recall only one row.

   • Result: They could recall almost 100% of any row indicated.

Conclusion: This proved that for a brief moment, their iconic memory held all 12 letters (large capacity). The "whole report" failed because the memory decayed (short duration) before they could "read out" all the letters from their internal icon.

Echoic Memory (Auditory Sensory Memory)

• Capacity: Large, can hold several pieces of auditory information.

• Duration: Significantly longer than iconic memory, lasting for several seconds (e.g., 2-4 seconds).

• Format: Raw, "unprocessed" phonological (sound-based) information.

• Purpose: This longer duration is crucial for understanding spoken language, allowing us to connect the end of a sentence to its beginning.

Short Term Memory 

• Capacity (Limit): Limited. 

• Duration (Forgetting): Short, decaying in approximately 20-30 seconds if maintenance rehearsal is prevented.

• Format (Code): Primarily phonological (sound-based). This is why we get acoustic confusion errors (e.g., mixing up "F" with "S") even when items are presented visually.

Chunking and Limit of ST Memory

The "7 +/- 2" limit of STM is not 7 items, but 7 meaningful chunks.

Chunking is the process of grouping individual, unrelated items into a single, larger, meaningful unit. This "works around" the limited capacity.

• Example: The letter string F B I C I A A B C is 9 items (at the limit of STM).

• Chunked: FBI CIA ABC is 3 chunks, which is easy to hold in STM.

• Hierarchical Chunking: You can group chunks into even larger chunks (e.g., grouping phone numbers by area code, then prefix).

Serial Position Effect + Primary and Recency Effects

When given a list of items to remember, recall is best for items at the beginning and end of the list, forming a U-shaped curve.

1. Primacy Effect: The superior recall for items presented at the beginning of the list.

2. Recency Effect: The superior recall for items presented at the end of the list.

Why are the Primacy and Recency effects thought to have different origins (i.e., from LTM and STM)

They are believed to be separate because you can manipulate them independently.

• Primacy (LTM): Items at the beginning get the most rehearsal, allowing them to be transferred to Long-Term Memory (LTM). Slowing down the list presentation improves the primacy effect (more time to rehearse).

• Recency (STM): Items at the end are still "fresh" in the Short-Term Memory (STM) buffer and are "dumped" out first.

• Evidence: If you add a distractor task (like counting backwards for 30 seconds) after the list, it prevents rehearsal and "wipes" the STM buffer. This eliminates the Recency effect but leaves the Primacy effect intact, proving they rely on different memory systems.

Amnesia of Clive Wearing

Clive Wearing suffers from a profound case of retrograde (loss of old memories) and anterograde amnesia (inability to form new memories) due to viral encephalitis that destroyed his hippocampus.

• His sensory memory and STM are intact, but he cannot transfer any new information from STM to LTM.

• His existence is described as a "moment-to-moment" consciousness. He feels as though he is "waking up" for the first time every 30 seconds, as his STM buffer decays and nothing is saved.

• His case provides powerful neuropsychological evidence for the separation of STM and LTM systems.

Working Memory

Working Memory (e.g., Baddeley's model) is a more modern, active, and complex representation of short-term storage.

• STM (Traditional): A single, passive "box" or "buffer" used only for temporary storage.

• Working Memory (Modern): An active "mental workbench" with multiple components. It's used for both storing and manipulating information to perform complex cognitive tasks (like reasoning, comprehension, and problem-solving).

3 Original Components of Baddeley’s Working Memory Model 

• Central Executive: The "manager" of the system. It controls, monitors, and allocates attention. It decides which information to process and which "slave systems" to use.

• Phonological Loop: A "slave system" that processes and stores auditory/verbal information (like a mental "inner voice" and "inner ear").

• Visuospatial Sketchpad: A "slave system" that processes and stores visual and spatial information (like a mental "inner eye"). (Note: An "Episodic Buffer" was added later to integrate information and link to LTM).

Evidence Supporting Existence of Phonological Loop 

• Phonological Similarity Effect: Recall is worse for lists of items that sound similar (e.g., F, S, X, M) than for items that sound different (e.g., F, T, R, K). This shows the loop stores info based on sound, not visual shape.

• Word-Length Effect / Speed of Speech: Your immediate memory span (digit span) is strongly correlated with how fast you can speak. You can remember more short words (e.g., "dog," "cat") than long words (e.g., "university," "hippopotamus") because the loop's "tape" is time-limited (about 2 seconds), not item-limited.

• Articulatory Suppression: If you are forced to repeat an irrelevant sound (like "the, the, the...") while learning a list, your recall performance plummets. This is because the articulation "blocks" the phonological loop, preventing you from rehearsing the new information.