PSC134-L17

📗 -> Lecture Date: Name

---

Lecture Slide Link

🎤 Vocab

❗ Unit and Larger Context

Active Maintenance can do a lot

• Active maintenance can do a lot!
  • Developments in this can explain how kids go from perseverating to successful switching in A-not-B and card sorting
  • Gradual changes can explain why kids (and frontal patients) show dissociations, seeming to know what to do but perseverating
  • Explains how adults succeed in the face of conflict (like in Stroop), but are slowed
• But how do we know what to maintain? (BG)

Dopamine driven learning in PFC-BG (Go and No Go) guides what to maintain, helps get rid of the homunculus

✒️ -> Scratch Notes

Active Maintenance can do a lot

Dopamine driven learning in PFC-BG (Go and No Go) guides what to maintain, helps get rid of the homunculus

Why does A not B model have bias?

• You would need to be able to maintain the toys location in order to succeed at the task (need strong WM).
  • Why is a certain amount of WM enough for A but not B?
    • Partially a capacity thing, wants to hold onto A
    • Needs more flexibility

  • Mechanism:

• Difference in weight training ofc, rewards strengthen connections.
  • ‘Listening’ more to A
• At first, the weak WM activation of seeing it hidden at A is enough to reach at A.
• Afterwards, the weak activity at B has to override the pull to reach/maintain A

Card Sort Task

A paradigm where they have two forms of sorting, color or shape.
Kids doing this will get the first rule right, but be unable to switch when the game changes.

• Representing one type of feature more actively.
  • One aspect has a stronger association of value
    • Reflect shortcuts in thinking
      When they’re given cards, the latent activations of card color is triggered.
• If they were given a card neither blue nor red they would get it right.
• They fail verbal questions as well, not just an action problem, a knowledge problem

🧪 -> Example

Describe to the person on the bus how the frontal cortex knows which goal-relevant information to actively maintain in working memory. Your answer should include information about how the basal ganglia supports a neural network model’s performance of the SIR task.

Our brain maintains working memory in the prefrontal cortex, and this info is used to execute everything from simple to difficult tasks. We learn what and how to maintain information through interactions between different regions of our brains, namely dopamine neuron firing and the basal ganglia. Our basal ganglia controls what enters our PFC through its Go and No Go pathways, and the dopamine firing is a reward driven process in which the basal ganglia learns how to best control its Go and No Go pathways. In the SIR (Store, Ignore, or Recall) task, we have to learn what is important to maintain and how to do it. Through trial, error, and reward, our basal ganglia will eventually find a strategy of how to pass or not pass information to our working memory (PFC) and allow us to perform the task. 

🔗 -> Links

Resources

• Put useful links here

Connections

• Link all related words