N8 - Stm/working Memory

Front 1

What is working memory? Not a model, network, process, formation?

Back 1

Working memory is a conceptual framework links to LTM, perception and action theory

Front 2

What are the purposes of the frontal lobe? How is it connected that might make its job plausible?

Back 2

• Motor planning/output
• Inhibiting activities
• Planning
• Holding info
• Changing line of thinking
  
• Monitoring actions
• Inhibiting ineffective plans

i.e. executive functions

Connects to thalamus with white matter connections - gets sensory info + lots of cortico-cortical white matter connections. White matter connections are faster, than regular cortical connections ∵ insulated with myelin

Front 3

What's the difference between short term memory and working memory

Back 3

Short term memory has a chunk limit (test by asking to determine if coloured squares presented before break are same as after - bigger set = more difficult) and temporal decay. Basically a limited store of information that eventually rots.

So what is working memory? Seems to be STM applied to tasks, holding/manip info in STM, using attention to manage STM∴ all about manipping STM.

Thing is both have evolved, so now:

• Sensory buffers are different - preconscious
• STM seems to be about visuospatial sketchpad, verbal rehearsal.
  
• WM seems to be about memory that can then be manipulated/acted on

Front 4

What are the models for STM

Back 4

Atkinson & Shiffrin (1968) model describe it in serial way. BUT STM / LTM dissociate (KF vs HM). + not just rehearsal to get from STM / LTM, also about depth of processing

Baddeley & Hitch (1974) - visuospatial sketchpad / phonological loop / episodic buffer (added later) (although may be that episodic buffer is above those two - may be working memory→ feed into central exec). Can distinguish those by asking them to do multiple tasks, and seeing which interfere e.g. digit memory + draw house is fine, but digit memory + transcribe tape is difficult.

Front 5

What are the models for Working Memory / Conscious processing

Back 5

• Norman Shallice (1980) - involves Supervisional attentional regulation by PFC of low level sensory info
• Baars (1989) - conscious info has to get onto global workspace for conscious knowledge / pass on to others
• Dehaene et al (1998) - global neuronal workspace→ high level processes e.g. attention, memory, associative sensory/motor → broadcast to lower levels influence what is picked into consciousness. Although low level can initiate its own stable cell assembly frontally i.e. break into consciousness
• Modern model is kind of an amalgamation of these - sensory buffers, with central executive that gives attention which biases what info is selected to come into consciousness i.e. working storage, and then that can be acted on or put into STM and then learning makes that LTM.

Front 6

What is the likely neuroanatomy of short term memory working memory

Back 6

Anterior IT lesions worsen delay dependent deterioration, Dorsal PFC lesions worsen set size based deterioration in picking recognised object from previously presented set (Petrides 2000) = short term memory?

Dorsolateral PFC damage associated with deficits in manip of verbal / spatial knowledge (right dlPFC→ manip info in STM, left dlPFC→ manip info for other reasoning contexts?) (Barbey et al 2012) = working memory?

Ask people to remember either colour or motion of blocks, then TMS right IFJ PFC→ decline in performance. IFJ PFC about biasing attention? Can be shown to lower EEG activity of visual cortex, so definitely makes an attentional difference in early visual areas (Zanto et al 2011)

Stroop task - activity Anterior Cingulate, Insula, Premotor, Inferior Frontal (Lamy et al 2000)

Metanalysis of Go-NoGo finds activation in pre-SMA and left fusiform, but complex tasks also have right prefrontal/parietal.

BUT is it likely to be loci - surely more network... graph of local/distal connections. Dual Task method shows large scale fronto-parietal activity... not just focal. (Sigman & Dehaene 2005)

Front 7

Can attention be controlled and which area might be responsible?

Back 7

Parahippocampal place area more active when asked to attend scenes and ignore faces AND reduced from passive levels if asked to ignore scenes (older patients struggle with suppression) (Gazzeley et al 2005)

TMSing IFJ PFC can be shown affect EEG activity of visual cortex, so definitely makes an attentional difference in early visual areas (Zanto et al 2011)

Matches up with Top-down attentional bias in biased competition theory.

Front 8

How can you test if the Central executive is the processing bottleneck?

Back 8

Use dual task method with two tasks that use diff sensory buffers (i.e. number comparison vs tone discrim) so that they both get to central executive at the same time→ see if it's a bottleneck.

Present one task slightly after the other - see if the delay actually affects RT, or if changing the task components makes a difference.

Prediction is that RT for Task1 is independent of delay for first task∵ it gets to exec first always, whereas for Task2 RT doesn't change for short delays, but eventually does, because always some delay while exec processing first task (interference). WORKS

Then change task, so perception takes longer (e.g. words vs. digits) or computation takes longer (similar vs. v different numbers) or Motor takes longer (two taps vs one tap). Can change either Task1 or Task2 (just present number discrim first or - see how that affects graph. Effects work as you'd predict in bottleneck model.

(Sigman & Dehaene 2005)

Front 9

How does the bottleneck theory fit in with Baar's global workspace theory?

Back 9

Baar suggests that specialised systems can do be active and perform tasks without conscious control or pool resources to trigger consciousness and enter global workspace→ site of information exchange i.e. now available for all other systems = "consciousness" = "working memory"?

Front 10

How was Baar's original global workspace theory built upon?

Back 10

Dehaene modified it from conscious vs. multiple unconscious competition to...

1. Subliminal (unattended) - not enough activity to trigger
2. Subliminal (attended) - feed forward activation
3. Preconscious - enough activation but top down attention prevents trigger
4. Conscious - enters global workspace because enough

Front 11

What experiments show ignition of consciousness?

Back 11

Using masks/attentional manipulation shows different levels of activation i.e. masked + unattended < masked attended < unmasked unattended < unmasked attended (Dehaene and Changeaux 2005)

Time for overlapping stim can be found by flashing stim then delay then mask, and at <80ms delay → maks lowers performance (objectively and subjectively). Can also look at which areas are affected by the mask → early visual areas unaffected by mask, higher visual areas effects vary by mask strength, whereas prefrontal only active with low mask. (Del Cul et al 2007)

Masking (tiny delay post stim) vs non mask → nonconscious vs. conscious activation→ Looked at EEG shows that both have early occipital, then later only unmasked i.e. conscious has anterior region (fusiform) activation = feedforward activation of frontal if stim big enough (Gallard et al 2009)

Similarly in Attentional blink paradigm→ some can see / some can't see same physical stim. Diff activation i.e. if you can see → inferior frontal activation (Sergent et al 2005)

Front 12

Can consciousness be explicitly studied

Back 12

Yes if the definition is restricted to kind of like attention but slightly different i.e. relies on bistable switch of consciousness which can be triggered by other modules (whose power can be amplified by attention). Basically global workspace or working memory type definition

Front 13

Are there constraints on contents / computations of consciousness

Back 13

Probably - based on structure / connectivity of the frontal lobe workspace

Front 14

What is the metacognition executive function and where is it based?

Back 14

Ability to dissociate effort and feeling of effort

Examples

• Left medial frontal cortex damage does adapt to more difficult stroop task, but judges them equally (Dehaene et al 2005)
  
• Asked to rate confidence in deciding if noisy stim is house or face - most people's accuracy correlates with confidence. Generally activates dACC, rlPFC more if has to assess confidence (vs moving mouse on a scale not related to accuracy). More activity in those areas if less confident (because more effort assessing?). More activity rlPFC is proportional to accuracy, too (i.e. activity → effort assessing→ better results)? (Fleming et al 2012)

Front 15

What conclusions can be drawn from (Canales-Johnson et al 2015)

Back 15

Task

1. Tap with external heartbeat
2. Tap with own heartbeat
3. Hear auditory feedback
4. Tap with own heartbeat postfeedback
5. Asked how well they think they learned
   

Conclusions

• Those good at interoception→ higher frontal activity (matches Fleming conclusion)
• Those that learn have increased right medial frontal lobe activity after feedback.
• Those that learn also have increased right frontal opercula activity - kind of superior insula (insula highly implicated in interoceptive awareness - gets info) so if you learn→ more weight on insular input
• If subjective assessment of learning ability correct→ more gamma phase synchrony post feedback - accuracy of metacognition reflected in gamma phase synchrony→ indicates communication between interoception (insula) and metacognitive areas (frontal). Similar to binding of attention