WORKING MEMORY
- Early theories for short-term memory (STM)
- Waugh and Norman's (1965) two-store model
- Atkinson and Shiffrin's (1968) model
- Critics of the two-store model: STM was considered as a unitary system
- Tracking mental work on-line
- single cell activity: Fuster and Jervey (1981), see Fig. 3-1, 3-2
- delayed matching-to-sample paradigm
- recorded from inferior temporal cortex of monkeys
- found increased electrical activity in specific neurons implicated in visual processing
- interpreted the increased activity as a reflection of mental work necessary to retain the
information of the sample stimulus
- Multiple-component model of working memory: Baddeley's working memory model
- Definition: the term working memory refers to a system that has evolved for the short term
maintenance and manipulation of information necessary for the performance of such complex tasks
as learning, comprehension and reasoning.
- three components, each fulfilling specific functions, see Fig. 3-3
- phonological loop
- concerned with auditory and speech-based information
- visuospatial sketchpad
- maintains and manipulates visual and spatial information
- central executives
- Evidence comes from different lines of research
- case studies
- dissociations observed in dual-task studies
- neuroimaging research
- Case KF
- Described by Warrington and Shallice (1969)
- A victim of a motorcycle accident
- injury to the left parieto-occipital region
- Severe limitations of verbal STM
- No recency effect in free recall learning
- Could not remember > 2 digits
- HOWEVER, able to acquire new information for the long term
- Experimental dissociation: Lee R. Brooks (1968), see Fig 3-4, 3-5
- Rationale of the duel task technique
- two tasks to be executed concurrently
- if performance suffers, then the two tasks engage the same working memory component
- if performance remain intact, then the two tasks depend on different working memory
components
- 2 x 2 factorial design: covaried the task (visual vs. verbal) and mode of response (visual vs. verbal)
- Visual task: visual imagery, Ss were asked to visualize a block letter, trace the letter mentally and
indicate whether the letter has an outside or inside corner.
- Verbal task:
- memorizing sentences such as "A bird in the hand is worth two in the bush"
- think of the sentence word by word and indicate for each whether it was a noun.
- Found dissociation between the two tasks, suggesting two separate components of working memory
- the visual response was faster in the verbal task
- the verbal response was faster in the visual task
- Support the distinction between the phonological component and the visuospatial sketchpad.
- Neuroimaging studies for working memory
- Delayed response paradigm combined with neuroimaging or single-cell recordings.
- Friedman and Goldman-Rakic (1994), spatial task
- Recorded blood-flow activity from monkeys trained to remember the location of stimuli on
the screen for a brief interval.
- During the interval, PET activity increased in the prefrontal cortex and inferior parietal
cortex.
- The greater the accuracy of the monkeys in remembering the stimuli, the greater was the
brain activity
- Cohen et al (1994), nonspatial task
- Human learners monitored letter sequences on a screen
- Press a key when a target letter was repeated, e.g., as in MXM as compared to MXB
- Prefrontal cortex became active as the Ss worked on the task
- The more the number of intervening letters, the larger area of the cortex was activated
- Jonides (1995)'s PET study on phonological buffer and visuospatial sketchpad, see Fig 3-6
- Two-back task (a phonological task)
- Ss see a sequence of individual letters (e.g. M P F P...) presented at a rate of one letter every
3 sec
- Ss make a yes response whenever a letter is shown that had occurred two positions back
- otherwise, no is the correct response.
- Dot-location task (a visual spatial task)
- presentation of three dots in different locations on screen for 200 msec
- followed by a blank screen for 3 sec
- test: an outline circle is presented on a location of the screen
- Ss must indicate whether the location was occupied by one of the dots of the previous screen.
- Neuroimaging revealed that
- the verbal task activated in several regions in the left hemisphere, in the frontal lobe, Broca's
area, the parietal lobe
- the spatial task activated in several regions in the right hemisphere, in the frontal lobe, the
parietal lobe and the occipital lobe.
- the PET data for the two tasks present a "clear dissociation between working memory for
phonological and spatial information.
- Spatial vs. visual processes in visual spatial sketchpad, see Fig 3-7
- Smith and Jonides (1997): separate spatial and visual subsystems in working memory
- The target stimuli were two irregular objects located in random locations of the screen
- The probe stimulus consisted of a single object
- In spatial memory task, the Ss were asked whether the probe was in the same location as one
of the target stimuli
- In visual memory task, the Ss were asked whether the probe had the same form as one of the
target stimuli
- Although the stimuli and trial sequence were identical in the two tasks, neuroimaging results
revealed a clear dissociation
- Given the spatial memory instructions, it was the right hemisphere that became active, esp.
the prefrontal cortex, the premotor cortex, the occipital cortex, and the parietal cortex.
- Given the visual memory instructions, two regions in the left hemisphere were activated: the
parietal cortex and the inferotemporal cortex
- The Central Executive
- Frontal lobe patients suffer from dysexecutive syndrome, this condition involves reduced control of
behaviour and difficulty in coordinating action to meet a specific goal
- Wisconsin card sorting task, the patients are said to be susceptible to proactive interference, the
inhibiting effects of prior learning on new learning
- angulate cingulate gyrus and stroop effect