Psych 240 Lecture 10 (2/7)

Why is human memory divided into multiple systems?

Memory is not one unified store. Different systems vary in capacity, duration, format/organization, and neural hardware. The major divisions include sensory memory, working memory, and long-term memory (explicit and implicit).

What are the four defining features used to compare memory systems?

Capacity: how much information can be held. Duration: how long information remains available. Format and organization: how information is represented and structured. Hardware: the neural regions supporting the system.

What are encoding, storage, maintenance, and retrieval?

Encoding transforms incoming information into a neural representation. Storage retains information over time. Maintenance keeps information temporarily available for ongoing processing, particularly in working memory. Retrieval reactivates stored information when needed.

What is sensory memory, and what are its two major forms?

Sensory memory briefly preserves perceptual information after a stimulus disappears. Iconic memory stores visual information; echoic memory stores auditory information. Both operate automatically and before full recognition.

What is iconic memory, and what does the fading icon demonstrate?

Iconic memory is a brief, high-capacity visual representation that persists after a visual stimulus disappears. A briefly flashed image appears to fade even after it is physically gone, demonstrating persistence of a visual sensory trace rather than continued stimulus presentation.

What are the capacity, duration, format, and hardware of iconic memory?

Capacity: extremely large, potentially nearly the entire visual input. Duration: approximately 150–300 ms after stimulus offset, with rapid fading. Format: visual/perceptual and pre-recognition. Hardware: early visual cortex, particularly V1.

Why was iconic memory difficult to investigate experimentally?

The visual trace disappears so quickly that participants cannot report all of its contents before it fades. Poor whole-report performance might reflect limited time or working memory capacity rather than a lack of information in iconic memory.

What is the whole-report paradigm, and what limitation does it reveal?

Participants briefly see an array of letters or numbers and attempt to report every item. They typically report only about 4–5 items, even when more were presented. However, this does not prove that only 4–5 items were initially stored because reporting takes time and working memory is limited.

Explain Sperling's (1960) partial-report experiment, including its procedure.

Participants briefly viewed a 3-row array of letters or numbers. After the array disappeared, a high, medium, or low tone indicated which row to report. Because the cue came after presentation, participants could not know beforehand which row would be tested.

Why did Sperling use a tone to cue only one row after the array disappeared?

The cue required participants to retrieve a subset from the briefly persisting visual trace. Reporting only four items avoided the whole-report bottleneck, while the post-stimulus cue ensured participants had to initially retain information from the entire array.

What were Sperling's main whole-report versus partial-report findings?

Whole-report performance plateaued at roughly 4–5 items. With an immediate partial-report cue, participants could report most items in whichever row was selected, implying that nearly the entire array had initially been available in iconic memory.

How is iconic memory capacity estimated from partial-report performance?

Estimated capacity = proportion correct in the cued row × total number of items in the array. For example, 95% correct with 12 total items gives 0.95 × 12 = 11.4 items. This assumes performance on the randomly cued row represents information available across the array.

What does the relationship between array size and partial-report performance reveal?

As array size increases, whole-report performance reaches a plateau, while estimated partial-report capacity continues increasing across the tested range. This suggests iconic memory initially holds far more information than can be transferred into working memory and reported.

What did Sperling discover about the format of iconic memory?

Participants could use cues based on perceptual features such as spatial location, but not effectively select items based on abstract categories such as reporting only digits. This supports the conclusion that iconic memory is perceptual and pre-recognition rather than fully semantic.

How did Sperling investigate the duration of iconic memory?

He varied the delay between the disappearance of the letter array and the cue tone. Immediate cues produced high partial-report performance, but increasing the delay reduced the number of items participants could retrieve.

What were Sperling's conclusions about the duration of iconic memory?

Iconic memory decays extremely rapidly: approximately half the information fades within about 100 ms, and the trace is almost entirely gone within roughly 300 ms. Delayed partial-report performance eventually approaches the whole-report level.

Why does a delayed cue reduce partial-report performance?

The iconic trace fades automatically after the stimulus disappears. If the cue arrives too late, information from the selected row is no longer available to transfer into working memory, even though it was initially present.

What were the three major conclusions of Sperling's iconic memory research?

1. Iconic sensory memory exists. 2. Its capacity is extremely large, potentially containing almost the entire visual input. 3. Its contents decay rapidly within a few hundred milliseconds.

What did Teeuwen et al. (2021) find about the neural basis of iconic memory?

Single-unit recordings in macaque V1 showed neural responses that decayed on a time course consistent with iconic memory. Stronger V1 activity was associated with better performance on a partial-report-like task, supporting early visual cortex as a neural basis for the icon.

What is echoic memory, and how does it differ from iconic memory?

Echoic memory briefly preserves auditory perceptual information after a sound ends. It lasts approximately 2–4 seconds, much longer than iconic memory's few hundred milliseconds. Both are automatic, pre-attentive sensory stores.

How can echoic memory complicate the interpretation of dichotic listening experiments?

Participants may report information from an unattended ear because some information passed through the attentional filter, or because attention shifted afterward to an echoic trace that was still available. The observation alone cannot distinguish these explanations.

What are the defining features of sensory memory overall?

Capacity: very large, especially for visual input. Duration: iconic approximately 150–300 ms; echoic approximately 2–4 seconds. Format: modality-specific, perceptual, and pre-recognition. Hardware: early sensory regions such as V1 for visual information.

What is short-term memory (STM), and how does it differ from sensory memory?

STM is the brief, relatively passive storage of a small amount of information. Unlike sensory memory, which retains large amounts of raw perceptual information for very short periods, STM has a sharply limited capacity and can maintain selected information longer.

What is memory span, and how is it measured with the digit-span task?

Memory span is the number of items someone can temporarily retain. In digit span, participants hear or see increasingly long sequences and reproduce them in order. Average performance in the lecture is approximately 7 ± 2 verbal items, although capacity depends on the material.

What is Miller's 7 ± 2 principle, and what is its limitation?

People can often retain approximately 5–9 meaningful verbal items or chunks in simple short-term span tasks. However, this is not a universal capacity limit: visual content often has a smaller capacity of approximately 2–5 items, and meaningful grouping changes apparent span.

What is chunking, and how does it improve short-term memory performance?

Chunking groups separate elements into meaningful units using existing knowledge. For example, FBI CIA NASA is easier to remember as three familiar abbreviations than as nine unrelated letters. It improves effective span without necessarily increasing the underlying storage capacity.

Explain Peterson and Peterson's (1959) short-term memory experiment.

Participants remembered a short letter sequence and then counted backward by threes from a number during a variable delay. Counting prevented rehearsal. Recall became worse as the delay increased, demonstrating rapid loss of unrehearsed short-term information.

What does the Peterson and Peterson study suggest about rehearsal and STM duration?

When rehearsal is prevented, short-term recall declines quickly over delays of several seconds. This shows the importance of maintenance for temporary retention, although the decline can involve interference as well as decay.

What is the serial position curve?

A graph showing recall accuracy as a function of an item's position in a studied list. It typically shows high recall for early items (primacy effect), lower recall for middle items, and high recall for final items (recency effect).

What is the primacy effect, and how is it explained by the modal model?

The primacy effect is better recall for items at the beginning of a list. The modal model explains it by proposing that early items receive more rehearsal and are therefore more likely to enter long-term memory.

What is the recency effect, and how is it explained by the modal model?

The recency effect is better recall for items at the end of a list. The modal model explains it by proposing that recent items remain in short-term memory at the time of immediate recall.

What happens to primacy and recency when a counting task is inserted before recall?

The recency effect largely disappears because the counting task disrupts the short-term store. The primacy effect remains because earlier items are thought to be supported by long-term memory. This dissociation supports separate short-term and long-term contributions to recall.

What is the Atkinson and Shiffrin (1968) modal model of memory?

A model proposing that information passes through sensory memory, then a single short-term store, and eventually into long-term memory. Rehearsal maintains information in STM and was proposed to facilitate its transfer into LTM.

What are the three major assumptions of the modal model?

1. More time in STM through rehearsal produces better transfer into LTM. 2. Information must pass serially from sensory memory through STM to reach LTM. 3. STM is a single, unitary store serving as the workspace for cognitive tasks.

What is the first major problem with the modal model?

1. Maintenance rehearsal does not reliably improve long-term storage. Simply repeating a phone number may keep it available temporarily without making it memorable the next day. The way information is processed matters more than time spent repeating it.

What is the difference between maintenance and elaborative rehearsal?

Maintenance rehearsal involves repeating information to keep it temporarily available. Elaborative rehearsal involves processing meaning and connecting new information to existing knowledge. Elaborative processing is more beneficial for long-term retention.

What is the second major problem with the modal model?

Some patients have severely impaired STM but relatively preserved ability to form new long-term memories. If STM were the mandatory gateway to LTM, damage to STM should prevent normal LTM formation, but this prediction is contradicted.

What did patient KF demonstrate about short-term and long-term memory?

Patient KF had severely impaired short-term verbal memory, including a digit span of approximately 1–2 items, while important long-term memory abilities remained relatively intact. This neuropsychological dissociation challenges the idea that all information must pass through an intact STM store to reach LTM.

What is a neuropsychological dissociation, and why is it important for memory models?

A dissociation occurs when brain damage disrupts one cognitive ability while leaving another relatively preserved. KF's impaired STM but preserved LTM suggests that the two systems are at least partly independent rather than one being an obligatory gateway to the other.

What is the third major problem with the modal model?

People can maintain a substantial digit load while performing reasoning or comprehension tasks with relatively little impairment. If one unitary STM store supported all cognition, filling it should severely disrupt these tasks.

Explain Baddeley and Hitch's (1974) dual-task experiment and findings.

Participants maintained increasingly large digit sequences while simultaneously performing cognitive tasks such as reasoning. Performance became somewhat less efficient as load increased, but reasoning remained surprisingly effective even with a near-span digit load. This challenged a single shared STM workspace.

What did Baddeley and Hitch conclude from the dual-task findings?

Short-term retention and complex cognition cannot depend entirely on one undifferentiated storage system. Working memory must include partly separable components that can operate in parallel, alongside attentional coordination.

What are the three problems with the modal model, and what theory replaced it?

1. Maintenance rehearsal does not guarantee LTM learning. 2. Impaired STM can coexist with preserved LTM. 3. Heavy digit loads do not eliminate reasoning ability. These findings motivated Baddeley and Hitch's multicomponent working memory model.

What is working memory, and how does it differ from traditional STM?

Working memory is a limited-capacity system that temporarily holds AND manipulates information during ongoing cognition. Traditional STM emphasizes passive storage, whereas working memory emphasizes active processing, attentional control, and multiple specialized components.

What are the four major components of the working memory model?

The central executive controls attention and coordinates processing. The phonological loop maintains verbal/speech-based information. The visuospatial sketchpad maintains visual and spatial information. The episodic buffer integrates information across subsystems and long-term memory.

What is the central executive, and what are its main functions?

The central executive is an attentional control system rather than a storage buffer. It directs attention to relevant information, coordinates working memory subsystems, switches between tasks, and retrieves information from long-term memory.

What are the three storage subsystems of working memory, and what does each store?

Phonological loop: speech-based/verbal information. Visuospatial sketchpad: visual appearance and spatial locations. Episodic buffer: integrated, meaningful chunks combining information from different working memory systems and long-term memory.

What are the two components of the phonological loop?

The phonological store, or inner ear, passively retains speech-based information for about 2 seconds. The articulatory rehearsal process, or inner voice, refreshes the store through subvocal repetition and can recode visually presented words into a phonological representation.

How does the phonological loop maintain visually presented verbal information?

Written words or letters can be converted into speech-based representations through subvocal articulation. The articulatory rehearsal process then repeatedly refreshes those representations before they fade from the phonological store.

What is the phonological similarity effect?

Immediate verbal recall is worse for letters or words that sound alike than for items with distinct sounds. For example, B, C, P, T, and V are more confusable than acoustically distinct letters because the phonological store represents sound-based information.

What did Conrad's phonological similarity findings demonstrate?

Participants made more short-term recall errors for acoustically similar letters or words. Similarity of meaning or visual appearance did not produce the same pattern. This supports phonological rather than primarily semantic or visual coding in the verbal short-term store.

Why do similar-sounding words interfere with recall more than similar-meaning words?

The phonological store represents verbal material in a sound-based format. Items with overlapping sound patterns become difficult to distinguish, whereas similar meanings alone do not produce the same acoustic confusion.

What is the word length effect?

Short words are generally remembered better than long words in immediate verbal span tasks. Longer words take more time to pronounce and rehearse, so fewer can be refreshed before the phonological trace fades.

What does the approximately 2-second capacity of the phonological loop mean?

Verbal span is constrained partly by how much speech can be articulated or refreshed in about 2 seconds. People can usually retain more rapidly spoken short words than slowly spoken long words, making pronunciation time important.

What did Ellis and Hennelly (1980) find about bilingual verbal memory span?

Welsh-English bilingual participants showed a larger digit span in English than in Welsh, despite speaking both languages fluently. This was linked to differences in pronunciation duration and supported the word length effect and articulatory rehearsal account.

What is articulatory suppression, and how is it performed?

Articulatory suppression involves repeatedly speaking an irrelevant word or syllable, such as 'the, the, the,' while trying to remember verbal information. This occupies the articulatory rehearsal process and prevents normal subvocal refreshing.

What are the main effects of articulatory suppression on verbal working memory?

It generally reduces recall accuracy by preventing rehearsal. For visually presented verbal material, it can eliminate the usual phonological similarity advantage for dissimilar items because visual words can no longer be recoded and rehearsed normally.

Why does articulatory suppression eliminate the phonological similarity effect for visually presented items?

Normally, visual letters are converted into phonological representations through articulatory rehearsal, creating sound-based confusion. Suppression blocks this recoding process, so the usual difference between acoustically similar and dissimilar visual items can disappear.

How do phonological similarity, word length, and articulatory suppression jointly support the phonological loop?

Phonological similarity shows verbal information is represented by sound. Word length shows retention depends on articulation time. Articulatory suppression shows that rehearsal is an active process needed to refresh or recode verbal information. Together they support distinct storage and rehearsal mechanisms.

What is the visuospatial sketchpad, and what information does it maintain?

The visuospatial sketchpad temporarily maintains visual and spatial information, including object shape, color, appearance, and location. The lecture describes an inner scribe that refreshes spatial representations by mentally tracing or placing objects within a visual cache.

What are the visual cache and inner scribe in the visuospatial sketchpad?

The visual cache maintains information about appearance, such as shape and color. The inner scribe is a spatial rehearsal mechanism that refreshes representations of locations, arrangements, and movement, such as mentally tracing a shape.

Explain Brooks's (1968) experiment, including its tasks, responses, and dependent variable.

Participants performed either a visuospatial task (mentally tracing a block letter and judging its corners) or a verbal task (judging words in a remembered sentence). They responded either spatially by pointing or verbally by saying yes/no. The dependent variable was response completion time.

What were the results of Brooks's (1968) experiment?

The visuospatial task was much slower with spatial pointing than with verbal responses. The verbal task was much slower with spoken responses than with spatial pointing. Interference was greatest when the task and response used the same type of processing.

Why does Brooks's experiment support separate working memory subsystems?

If all temporary processing used one identical store, the interference pattern should not depend so strongly on task-response modality. The selective interference between spatial tasks and spatial responses, and between verbal tasks and vocal responses, supports distinct visuospatial and phonological systems.

What is the episodic buffer, and what is its approximate capacity?

The episodic buffer is a limited-capacity store of approximately four integrated chunks. It binds information from the phonological loop, visuospatial sketchpad, and long-term memory into unified, meaningful episodes or representations.

Why can people recall a meaningful 15-word sentence more easily than 15 unrelated words?

The episodic buffer can combine words with existing linguistic and semantic knowledge into meaningful phrases or chunks. A coherent sentence can therefore be maintained as a smaller number of integrated units, whereas unrelated words require many separate representations.

How is the episodic buffer different from the phonological loop and visuospatial sketchpad?

The phonological loop specializes in speech-based information, and the visuospatial sketchpad specializes in visual/spatial information. The episodic buffer integrates multiple formats and links working memory contents with long-term knowledge into coherent chunks.

What does Baddeley's reading argue about working memory capacity and intelligence?

Working memory span measures involving both processing and storage are associated with performance in reasoning and comprehension. However, working memory capacity is not identical to intelligence, and correlations do not mean the two constructs are the same.

What is a complex working memory span task, and how does it differ from simple digit span?

A simple digit-span task mainly tests temporary retention. A complex span task combines processing with storage, such as reading sentences while remembering their final words. It better captures the demands of maintaining information during ongoing cognitive activity.

What is the supervisory attentional system (SAS) discussed in Baddeley's reading?

The SAS is an attentional control framework used to explain nonroutine, novel, or conflicting situations requiring deliberate regulation rather than automatic habits. Baddeley relates it to the central executive, particularly in coordinating attention and controlling responses.

What is random generation, and why is it relevant to the central executive?

Random generation requires producing an unpredictable sequence of letters or numbers while avoiding familiar patterns and repetitions. It requires attentional control and becomes disrupted by competing executive tasks, making it useful for investigating central executive functions.

What does Baddeley's reading suggest about working memory and language learning?

The phonological loop may help people acquire new vocabulary by temporarily maintaining unfamiliar sound sequences. Evidence includes relationships between nonword repetition and vocabulary development, as well as difficulties learning new words when phonological short-term memory is impaired.

What is nonword repetition, and why is it useful in studying vocabulary acquisition?

Nonword repetition requires hearing and repeating an unfamiliar sound sequence. It tests the ability to temporarily encode and maintain new phonological patterns. Better performance is associated with stronger vocabulary learning, supporting a role for phonological working memory in language acquisition.

What is the overall defining profile of working memory according to Lecture 10?

Capacity: limited and task-dependent, including approximately 2 seconds of speech and roughly 7 verbal items in simple span tasks, with chunking improving performance. Duration: tens of seconds, with loss strongly affected by interference. Format: multiple verbal and visuospatial representations. Hardware: distributed systems, with further neural detail reserved for the next lecture.