Task 2: Cognition Flashcards
A1: Clarify the location/components of the lateral Prefrontal Cortex
located in the inferior frontal sulcus (IFS)
includes the inferior frontal junction (IFJ) & more anterior IFS
A1: How does the human lPFC map onto the macaque PFC
Macaque principal sulcus –> human IFS
Area 46:
- monkeys –> entire principle sulcus
- humans –> anterior part of middle frontal gyrus (kinda just above IFS)
A2: Clarify the interaction between DAN, the lPFC (central role) and medial action selection structures (Task 1)
(reference to Wong scheme)
DAN:
- frontoparietal attention system
- FEF: object selection
- lateral IPA: attention
- -> influence formation of motor goal
- part of Wong’s “what”
lPFC: Task Rules
- -> integrates sensory info from attention system
- -> affects object selection in FEF via task rules/goals
- -> biases object selection towards task-relevant features
- -> IFJ is invovled in flexibly switching between DAN & VAN
- part of Wong’s “what”
Pre(SMA), SC, vPreMC: Action Selection
- -> receive info on motor goal (which was generated by other 2 systems)
- -> Choose appropriate motions based on goal
- -> can bias attention to movement-relevant objects
- part of Wong’s “how”
A3: Why is the lPFC an “integrator of cognitive episodes”?
- temporal representations
lPFC has temporal representations “tags/place holders” of relevant task elements
- -> these are stored across cortex (language, sensory, motor, …)
- -> when goal is chosen, peripheral connections are activated/primed via BG (see task 8 BG biasing)
- -> these remote representations are maintained via feedback modulation
A3: Why is the lPFC an “integrator of cognitive episodes”?
- Baldauf & Desimone study on task-relevant synchrony (Xu article)
Stimuli: two overlapping streams of images (faces or houses) presented at slightly different timing (temporal frequencies)
Task: subjects either attended to houses or faces for 1-back task
Results:
FFA –> active for face attending
ParaHC Place Area (PPA) –> active for house attenting
IFJ –> active for any non-spatial attention
-> FFA & PPA were modulated by top-down attention according to the task demands (house vs. face)
- IFJ activity was synchronized with relevant sensory area for each task
- Phase-lags between IFJ and FFA/PPA (IFJ active ~20ms before FFA/PPA)
- -> IFJ drives synchonized activity with sensory areas (also anatomically connected to these areas)
- -> IFJ must bias perception through neural synchrony
A4: Explain Miller’s model of lPFC function
- Box 1 Figure
Network cues (C), responses (R) and hidden units (lPFC)
Same C1 could lead to either R1 or R2 depending on item information integrated in lPFC
E.g. Phone rings (C1) --> you are at home (C2) --> you answer (R1) Phone rings (C1) --> you are gone (C3) --> you don't answer (R2)
How does learning work in Miller’s model? Hint: rewards
Rewards
VTA (responds to reward first, later also just to cue)
–> sends DA to vmPFC & lPFC & affects plasticity there
–> stronger Cue-Reward connections –> learning
Learning -> Habituation -> Automaticity
- Reward signals –> strengthen PFC processing
- -> development of activity pattern reflecting association between goal-relevant info & situation
- -> after learning, even partial information can trigger whole representation & inform about appropriate response
- -> connections develop independent of lPFC
-> lPFC is important for initial learning of a rule
A4: Present evidence for Miller’s Model
- lPFC directly interconnected with higher-order sensory & motor cortex
- lPFC indirectly connected with limbic areas (via vmPFC) that process reward –> together: hidden unit
- many lPFC neurons show multimodal responses
- many lPFC neurons reflect learned cue-reward-action associations
- lPFC neurons show strong experience-dependent plasticity
Rule-training monkeys
- Same cue but 3 different rules
- -> PFC neurons selectively fired for each rule
A4: Miller’s model
1) Storage of elements from logical structure of the task
Task options: associative relationships between representations of different events/states/actions
–> PFC –> connections between cues (sensory) actions (motor) and rewards (limbic)
A4: Miller’s model
2) Sustained activity
What if distractors appear?
Task: monkeys retained task info over a delay period with distractors (cue –> delay –> sample –> response)
Sustained activity in PFC –> to maintain sample memory
Sustained activity in visual areas too –> BUT these were more easily disrupted by distractors!
A4: Miller’s model
3) Biasing remote areas
Biased competiton model
PFC –> top-down influence to prime task-relevant processing
- excitatory signal into posterior regions
Biased competition model
- -> Enhancement of activity in relevant neurons
- -> (Mutual) inhibition of irrelevant neurons
A4: Miller’s model
4) PFC Patients & switching flexibility
Wisconsin Card Sort Test (WCST)
PFC damage
- -> intact standard perception, memory & motor skills
- -> impaired ability to organize lives
- -> impulsitivty/irresponsibility–> stimulus-bound/cannot override impulses
- -> increased distractibility (loss of info maintenance during delay)
WCST
- -> intact learning of first rule
- -> inability to change to new rules
- -> missing flexibility to shift between rules & override pre-learned responses
A5: Why did people think the lPFC is involved in Working memory?
Goldman-Rakic (still pro-WM)
- PFC lesion –> impaired delayed response
- sustained activity during delay
- crucial role of PFC in learning rules
A5: Give evidence that lPFC is NOT involved in Working Memory (WM)
What makes activity in lPFC special
Monkeys vs. humans anatomy
- results from monkeys cannot be directly related onto A46/lPFC in human
- PFC not just active during WM but in other non-WM/cognitive tasks
- Delay activity is also seen in dorsal parietal cortex (dPC)
A5: What could sustained activity represent?
1) Maintenance of sensory information
2) Response preparation
3) Transformation
4) Task rules (Miller, Bunge)
5) Goals/rewards
