neuronal mechanisms of working memory

Question 1

what did Fuster 1974 do

Answer 1

monkeys see a piece of food in a tray, a shutter comes down and the tray is closed. Then when the shutter opens the monkey has to remember where the food was located.
Single neuron recordings from PFC showed elevated neuronal firing during the delay period, i.e. when the shutter is down.
This was interpreted as showing that neurons in the PFC hold a representation of the to-be-remembered stimulus (e.g. location of the food).

Question 2

what Monkey neurophysiology studies suggested a role for the prefrontal cortex (PFC) in working memory

Answer 2

fuster 1974

Question 3

who came up with Standard model of working memory

Answer 3

Goldman-Rakic (1987)

Question 4

what did Goldman-Rakic (1987) propose

Answer 4

In 1987, Patricia Goldman-Rakic proposed what’s become known as the standard model of WM, which suggests that this PFC activity reflects the neuronal instantiation of Baddeley’s working memory storage buffers.
(which contain a template, is a temporary representation of the information maintained during the delay).
This became the dominant view of the role of the PFC in working memory for some time.

Question 5

what did Funahashi 1989 do

Answer 5

In this task monkeys saw a cue on the left or right of fixation and had to maintain their eye gaze at the centre for 3 seconds and then make an eye movement (a saccade) in the direction of the cue. So they had to hold the direction of the cue in memory for 3 seconds.

They found that single neurons in PFC showed direction-specific firing during the delay period of this task (between the cue and response).
This particular neuron fired strongly to locations in the upper left quadrant.

Question 6

how was funashis results interprettred

Answer 6

They found that single neurons in PFC showed direction-specific firing during the delay period of this task (between the cue and response).
This particular neuron fired strongly to locations in the upper left quadrant.
They interpreted this as showing a direct neurophysiological correlate of a WM template, a temporary representation of the spatial location indicated by the cue.

Question 7

what study is Supporting human neuropsychological evidence for a role of the PFC in working memory

Answer 7

Petrides & Milner (1982)

Question 8

what did Petrides & Milner (1982) do

Answer 8

Self-ordered WM task given to patients with frontal and temporal cortex lesions.

Patients instructed to touch one picture per sheet of paper and not to touch the same picture twice.

Patients with frontal lesions disproportionately impaired.

There was also evidence for the standard model from human neuropsychology studies.
Petrides and Milner administered a task called a self-ordered task to patients with frontal lobe lesions, in which they saw sheets of paper like this.
The patients had to touch one of the images and then would be given the next sheet where they would have to touch a different image and so on until they had been through 12 sheets and touched all 12 images.
Thus, the task requires them to remember from one sheet of paper to the next which images they have touched previously.
They found patients with frontal lobe lesions were disproportionately impaired on this task indicating that patients had a deficit with WM.
Again they interpreted this as evidence that the PFC holds a representation of to-be-remembered information over the short periods of time of a WM task.

Question 9

what does ventral PFC do

Answer 9

object working memory (remembering what an item was)

Question 10

what is a key tenets of the standard model

Answer 10

key tenets of the standard model is that there are two types of visuospatial working memory – one system for objects and another for spatial locations.

Question 11

Evidence from monkey neurophysiology for what/where dissociation in working memory

Answer 11

Wilson et al 1993

Question 12

what did Wilson et al 1993 do

Answer 12

In this study monkeys were trained to perform an occulomotor delayed response task like the one seen previously.
Monkeys saw a cue that instructed them to make an eye movement in a particular direction, then had to remember that information before making the response.
The key change to the task was that in addition to the standard spatial cues (a cue appearing in the actual location to which monkeys had to make an eye movement) there were also pattern cues, in which a pattern appeared in the centre of the screen and the pattern instructed monkeys to make an eye movement in a particular direction. Thus, the response in both trial types was the same but the type of information to be remembered on each trial differed. In spatial trials the monkey had to remember the spatial location of the cue and in pattern trials the monkey had to maintain the identity of the pattern.

Question 13

what was Wilson et als results

Answer 13

the neuron in the inferior ventral PFC shows higher activation to the patterns and lower activation to spatial cues
Whereas the neuron in the upper dorsal PFC shows higher activation to spatial cues and lower activation to patterns.
You can also see that each neuron preferred a particular pattern or a particular direction

Question 14

what did Courtney et al 1996 do

Answer 14

Object WM task – remember the identities of 3 faces
- Activation in ventral PFC

Spatial WM task – remember the locations of 3 faces
- Activation in dorsal PFC

Question 15

Neuropsychological evidence for a role of the PFC in working memory

Answer 15

Petrides & Milner (1982)

Question 16

what did Petrides & Milner (1982) ACTUALLY do

Answer 16

• In this task, lots of other processes other than storage are actually required, which could have led to the deficit in patients.

What processes does this task require?

• Storage of previously touched item
• Suppression of previously touched item
• Selective attention to a novel item
• Planning/strategy use
• Sustained attention

Question 17

what did Courtney et al (1996) ACTUALLY do

Answer 17

Activation in dorsal PFC for spatial WM and ventral PFC for object WM
Possible confound – type of processing
- Subjects may use strategies (e.g. chunking) to help performance of the spatial task. This requires manipulation of stored information in WM

Perhaps it is this process of mental manipulation that drives activation in the dorsal PFC

Question 18

Perhaps it is this process of ________________ that drives activation in the dorsal PFC

Answer 18

Perhaps it is this process of mental manipulation that drives activation in the dorsal PFC

Question 19

Evidence against standard model – monkey neurophysiology

Answer 19

Rao et al. (1997)

Question 20

what did Rao et al. (1997) do

Answer 20

Task required monkey to remember an object and first make an eye movement to the correct object, then make an eye movement to the correct location
Over half of neurons showed both object and location selectivity
- Neurons in PFC adapt flexibly to represent task relevant information

Question 21

what did Rao et al find

Answer 21

This study found that single neurons in PFC can encode both the location of an object and also its identity.
Monkeys had to remember first the identity of an object and make an eye movement to the same object in a different location, and then remember its location, and they made an eye movement to the location lacking identity information.
Found that neurons could encode both object identity and location – suggests flexible adaptation of responses in PFC – neurons can adapt to represent whatever information is task relevant.

Question 22

What did Lebedev et al 2004 do

Answer 22

Monkeys saw a cue in a location, and the cue then moved. Then the cue either brightened in which case they made a saccade to the attended location, or it dimmed and they made a saccade to the remembered location

They labelled neurons in PFC as either attention selective or memory selective or hybrid.
Found that the majority of neurons were selective for the attended location, not the one held in memory.
Quite a few again showed a combination of responses, again consistent with the flexible adaptation theory.

Question 23

fMRI evidence for a ventral/dorsal dissocation in PFC according to type of processing

Answer 23

D’Esposito et al (1999)

Question 24

what did D’Esposito et al (1999) do

Answer 24

D’Esposito and colleagues provided fMRI evidence to support the idea that PFC is actually organised according to the type of processing carried out rather than the type of stimulus maintained.
They asked subjects to perform a WM task in which they had to either maintain information in WM by simply holding a letter string in memory and then judge whether a probe letter was part of the memory set. Or in the manipulation condition subjects had to rearrange the letters into alphabetical order and judge whether a probe letter was in a specific location in the letter string after rearranging.

Question 25

what did D’Esposito et al (1999) find

Answer 25

They found activation in both dorsal and ventral PFC during the delay period of the WM task.
However, as you can see here, activation in dorsal PFC was greater during the delay period of manipulation trials than during the delay period of maintenance trials
This suggests that the PFC is indeed organised according to the type of processing (dorsal = manipulation, ventral = maintenance) rather than the type of stimulus maintained (spatial vs nonspatial).

Question 26

• PFC may be organised according to the type of processing

Answer 26

– Dorsal: Manipulation

– Ventral: Maintenance

Question 27

why do we use Multivoxel pattern analysis

Answer 27

Multivoxel pattern analysis
• Takes advantage of fine-grained patterns of activation in the brain
• Uses machine learning techniques to teach an algorithm about the pattern of neural activation associated with a particular stimulus
• Algorithm is then able to ‘decode’ what the subject is looking at simply by viewing their pattern of brain activity
• Can we decode what item a subject is holding in WM from patterns of activation in the PFC?

Multivoxel pattern analysis takes advantage of the fine grained patterns of activation across groups of voxels.

Question 28

what did Linden et al 2011 do

Answer 28

• Subjects performed task requiring them to hold several objects in WM
• On each trial, required to decide whether a single object was part of the memory set
• 4 categories of objects – faces, bodies, flowers, scenes
• Trained pattern classifier to learn patterns of activation for each category
• Tested ability of pattern classifier to predict which category subject was holding in WM on each trial

Question 29

what did Linden et al 2011 find

Answer 29

As you can see, the regions holding category related information are exclusively in the posterior part of the brain. These are visual processing regions, for example the fusiform face area and other regions of the ventral visual stream – the same regions that are activated when these categories of object are presented to the subjects.
The implication of this is that the same regions that enable us to process an object when we see it with our eyes are also involved in storing temporary representations of those objects in WM.

Question 30

Can we decode information about items held in WM from patterns of activation in prefrontal/parietal cortex?-

Answer 30

Riggall & Postle 2012

Question 31

what did Riggall & Postle 2012 do

Answer 31

Riggall and Postle also looked into the question of where in the brain different types of information are represented during working memory.
They scanned people performing a WM task where they had to memorise a moving dot array, and during the delay period were cued to either remember the direction or the speed of the dots.
Then they were shown a probe that was either a match or mismatch.
The key thing here is that when the cue appeared subjects had to follow a rule (attend to the speed or the direction of the dots in WM) and also hold in WM a specific stimulus (the particular speed at which the dots were moving, or the direction in which they were moving).

Question 32

what did Riggall & Postle 2012 find

Answer 32

They found that they could decode which direction and how fast the dots were moving but only from visual cortex and temporal cortex. PFC provided no information about this.
In contrast, task instructions could be decoded from PFC and parietal regions.
This is consistent with the Linden study – stimulus specific information was stored in posterior, sensory specific cortex. However, information about task rules was stored in frontal and parietal cortex.

Question 33

An alternative view to the ‘standard’ model – WM as an emergent property

Answer 33

Postle (2006)

Question 34

what did Postle (2006) do

Answer 34

“If the brain can represent it, the brain can also demonstrate working memory for it.…working memory functions are produced when attention is directed to systems that have evolved to accomplish sensory-, representation-, or action-related functions. From this perspective, working memory may simply be a property that emerges from a nervous system that is capable of representing many different kinds of information, and that is endowed with flexibly deployable attention”.

Question 35

fMRI evidence for the internal attention hypothesis of WM-

Answer 35

Higo et al 2011

Question 36

what did Higo et al 2011 do

Answer 36

• Subjects maintained 2 objects in WM
• Subsequently cued either to maintain both objects (non-selective attention condition) or a single object (selective attention condition) in WM
• Finally asked to decide whether any of the objects in an array matched the object(s) they were holding in WM

A study by Higo and colleagues provided direct evidence for the idea that the PFC performs an attentional role in WM.
Subjects held two objects in WM and were subsequently cued either to maintain both objects (non-selective attention condition) or a single object (selective attention condition) in WM
Finally asked to decide whether any of the objects in an array matched the object(s) they were holding in WM

Question 37

what did Higo et al fnd

Answer 37

Activation in PFC was greater for the selective condition than the non-selective condition
Connectivity analysis demonstrated that activation in PFC modulated activation in different occipitotemporal regions depending on which stimulus the subject maintained in WM
Combined TMS/fMRI showed that disruption of PFC activity caused distal effects on occipitotemporal activation, thus establishing direction of causation
PFC seems to send an attentional bias signal to sensory-specific regions to enhance processing of the task-relevant object during WM

They found that activation in inferior PFC was greater for the selective attention condition than the non-selective attention condition.
The implication is that the PFC performs an attentional function by prioritising processing of the item in WM that is most relevant to the current task.
They also found that this activation in PFC directly modulated activation in posterior regions, suggesting that the PFC biases activation in sensory-specific regions during WM in exactly the same way as it does during attention to external stimuli.
Thus, at least in this type of task, WM seems to involve a process of directing attention to specific items held in WM.

Question 38

• PFC neurons show sustained activity …

Answer 38

• PFC neurons show sustained activity during the delay period of a WM task
• However, this activity does not reflect storage of items in WM

Question 39

• PFC seems to play a role in …

Answer 39

• PFC seems to play a role in the processing of stimuli held in WM
• This role may involve enhancing attention to internal representations of task relevant stimuli in working memory and manipulating such information