PS425 Lecture 6: Working Memory Study Notes on Working Memory

Module Overview: PS425 / PS952 / EC956 Cognitive Psychology

• Module Coordinator: Prof Geoff Ward, who guides the overall structure and content of the course, ensuring alignment with current research and pedagogical practices.

• Core Areas: The module explores essential aspects of cognitive psychology, including:

  • Perception: The process by which sensory information is organized and interpreted.

  • Attention: Exploring how we allocate cognitive resources and filter information.

  • Memory: Investigating the encoding, storage, and retrieval processes of information.

  • Language: Examining the cognitive processes underlying language comprehension and production.

• Lecture focus: Lecture 6 delves into high-level concepts in memory, particularly the Modal Model developed by Atkinson and Shiffrin in 1968, which segregates memory into distinct stores, and the Working Memory Model proposed by Baddeley and Hitch in 1974, which offers a more nuanced understanding of short-term storage.

B: Classic studies of STM and the Modal Model

• The Modal Model (Atkinson and Shiffrin, 1968, 1971): This foundational theory posits the existence of three memory stores:

  • Sensory registers: Briefly hold sensory information (lasts seconds).

  • Short-Term Store (STS): A limited-capacity storage (usually about 7 ± 2 items, as noted by Miller) for immediate information.

  • Long-Term Store (LTS): A more permanent storage that can hold vast amounts of information for extended periods.
     

• Free Recall Signatures: The performance patterns observed in free recall conditions highlight key characteristics of memory retrieval:

  • Primacy Effect: Enhanced recall for early items due to extra rehearsal and successful transfer to long-term memory.

  • Asymptote: Refers to the tendency to recall middle items from LTS with diminishing returns as lists increase in length.

  • Recency Effect: Superior recall for final items, attributed to their availability in short-term storage at the time of retrieval.

• Key Evidence: Numerous studies bolster the Modal Model's validity:

  • Rundus (1971): Established a correlation between rehearsal frequency and the primacy effect, but noted no such correlation for the recency effect, emphasizing distinct storage processes.

  • Murdock (1962): Demonstrated that longer list lengths influence the success of recalling early and middle items but do not directly impact the recency effect.

  • Glanzer and Cunitz (1966): Showed that slower presentation rates enhance recall for LTS items; conversely, filled delays (like performing mental arithmetic) selectively disrupt the recency effect, indicating a contrast between STS and LTS.

• Neuropsychological Evidence: Baddeley and Warrington's (1970) study revealed that amnesic patients exhibited compromised LTS recall while still preserving the recency effect during immediate recall tasks, supporting the independence of STS and LTS functions.

C: Problems of the Modal Model of STM

• Rehearsal Discrepancy: Evidence (such as the penny study by Nickerson and Adams, 1979) indicates that increased rehearsal does not guarantee improved recall, challenging the idea of rehearsal being the sole mechanism for memory transfer.

• Long-Term Recency: Situations exist where recency effects manifest despite STS being theoretically empty, as evidenced in the Continual Distractor Task (Bjork and Whitten, 1974), and specific contextual tasks like car parking recall (da Costa Pinto and Baddeley, 1991).

• Temporal Contiguity: Research by Howard and Kahana (1999) illustrated that participants tend to recall items in the order presented, even after distractor tasks, pointing to underlying mechanisms in memory organization.

• Patient KF: Documented by Shallice and Warrington (1970), Patient KF exhibited significant impairment in STS (with a digit span of only 2-3) but showcased normal LTS learning, calling into question the premise that STS is a necessary precursor to LTS.

D: Contemporary Theories of Recency

• The Ratio Rule: Proposes that recency is proportional to the ratio $ rac{Dt}{T}$, suggesting that the timing of presentations greatly influences memory retention and recall.

• Theories: Current frameworks attempt to refine our understanding of recency effects:

  • Unitary Memory Accounts: These theories posit a single mechanism for recency phenomena, citing studies by Brown, Neath, and Chater (2007), as well as Howard and Kahana (2002).

  • Dual Store Accounts: Suggest separate mechanisms are involved in STS and LTS recency, with insights from Davelaar et al. (2005) and Unsworth and Engle (2007).

E: Active Maintenance and the Working Memory Model (WMM)

• Memory Span Definition: Refers to the longest list length that can be perfectly recalled in order 50% of the time (Guilford and Dallenbach, 1925), serving as a key measure for working memory capacity.

• Working Memory Model (Baddeley and Hitch, 1974): This model replaces the unitary STS concept with a framework comprising three specialized components:

  • Central Executive: Manages attention and allocates resources to tasks, effectively overseeing the flow of information.

  • Phonological Loop: Acts as a verbal STS that holds speech sounds and is sensitive to phonological similarity and articulatory suppression, with phenomena like the Word Length Effect significantly demonstrating its operation (Baddeley, Thomson, and Buchanan, 1975).

  • Visuo-Spatial Sketchpad (VSSP): Responsible for visual and spatial information processing, evidenced through tasks like the Brooks Matrix Task and Corsi Block Task.

• Evidence for Separation: Concurrent load tasks involving memory demands (like recalling 6 digits) reveal reduced performance on tasks involving reasoning and comprehension without affecting recency effects in free recall, thereby indicating a lack of trade-off among memory components.

• Patient ELD: Highlighted in a study by Hanley, Young, and Pearson (1991), Patient ELD exhibited intact verbal STS capabilities but struggled with visual-spatial tasks, showcasing the differentiation between verbal and visual memory stores.

F: Problems for the Working Memory Model

• Word Length Effect Validity: Diverse findings suggest that the effect relies on specific word sets; anomalies in studies (Lovatt et al., 2000; Caplan et al., 1992) show that longer words can sometimes be recalled effectively, complicating the straightforward interpretation of this phenomenon.

• Trace Decay: Research led by Vallar and Baddeley (1982) indicated that articulatory suppression does not universally lead to forgetting as decay theories would anticipate, challenging fundamental assumptions about memory loss processes.

• Limited Scope: The model does not sufficiently account for the Modality Effect, which indicates enhanced recency for auditory stimuli over visual ones, nor does it adequately explain how LTM can influence serial recall performance.

• Complex Span Tasks: The original WMM model struggles to encompass the challenges presented by complex span tasks, which require the retention of information while simultaneously processing other cognitive demands.

G: Current Accounts and Summary

• Embedded-Processes Model (Cowan): Proposes that working memory consists of the activated subset of LTM, emphasizing fluid interaction between short- and long-term storage.

• Time-Based Resource Sharing (TBRS) Model (Barrouillet and Camos): Suggests that separate stores maintain information through articulatory rehearsal supplemented by attentional refreshing mechanisms, accommodating the complexity of memory tasks.

• Summary: Contemporary theories advance beyond the foundational frameworks laid by Atkinson and Shiffrin and Baddeley and Hitch, proposing more dynamic models addressing crucial aspects such as long-term recency, attentional control, and the seamless integration between LTM and immediate memory processes.

- **Rehearsal Discrepancy:** Increased rehearsal doesn't guarantee improved recall, questioning its role in memory transfer (Nickerson and Adams, 1979). - **Long-Term Recency:** Recency effects can appear even when short-term store (STS) is theoretically empty (Bjork and Whitten, 1974). - **Temporal Contiguity:** Participants often recall items in the order presented, suggesting intricate memory organization mechanisms (Howard and Kahana, 1999). - **Patient KF:** Exhibited poor STS (digit span of 2-3) but normal long-term storage (LTS), challenging the necessity of STS for LTS learning. - **The Ratio Rule:** Proposes that recency relates to the timing of presentations, impacting memory retention. - **Theories of Recency:** - **Unitary Memory Accounts:** A single mechanism explains recency effects (Brown et al., 2007). - **Dual Store Accounts:** STS and LTS operate through separate mechanisms (Davaelar et al., 2005). - **Working Memory Model (WMM):** Consists of three components: - **Central Executive:** Manages attention and resources. - **Phonological Loop:** Handles verbal information, affected by phonological similarity (Word Length Effect). - **Visuo-Spatial Sketchpad:** Processes visual and spatial info (demonstrated via specific tasks). - **Patient ELD:** Showed preserved verbal STS but impaired visual-spatial tasks, indicating distinct memory systems. - **WMM Problems:** - **Word Length Effect:** Varies across word sets, complicating interpretations. - **Trace Decay:** Articulatory suppression doesn't always lead to forgetting, challenging decay theories (Vallar and Baddeley, 1982). - **Limited Scope:** Fails to explain the Modality Effect and its influence on recall. - **Complex Span Tasks:** Struggles with tasks requiring simultaneous information retention and processing. - **Current Accounts:** - **Embedded-Processes Model (Cowan):** Suggests working memory activates a subset of LTM, allowing interaction between the two. - **Time-Based Resource Sharing (TBRS) Model:** Maintains information through articulatory rehearsal and attentional mechanisms, explaining memory task complexity. - **Overall Summary:** Contemporary theories build on earlier frameworks by enhancing understanding of recency, attentional control, and integration of LTM with immediate memory processes.