Task 6: Error-monitoring

Question 1

A1: Discuss the neuranatomical organization of the mPFC

Answer 1

mPFC:

• vmPFC
  
  • > overlaps with mOFC
• dmPFC
  
  • > PreSMA & SMA
  • > medial part of superior frontal gyrus
• Cingluate cortex
  -> Gyrus vs. cortex (think of sulci)
  -> ACC (affective)
  pACC (pregenual)
  sACC (subgenual)
  -> refered to as rACC (Taylor)
  -> MCC (motor)
  aMCC
  pMCC
  -> referred to as dACC, ACC or rCMA

Question 2

A2: Explain the functional role of the motor section of the mPFC (think of MCC in Rushworth’s model)

Answer 2

• MCC -> action selection, action monitoring
• vmPFC -> outcome monitoring, reward-based decision-making/reward expectation (closely related to outcome monitoring)
• pACC -> social processing (self-knowledge, person perception, mentalizing)

Question 3

A2: Discuss the proposed functional division between MCC & SFG. Draw parallels to similar organizations of object selection (task 5) regions.

Answer 3

anterior/inferior MCC

• > action monitoring
• weighing/updating/evaluating action-outcome associations
• Task 5: lOFC for weighting/updating/evaluating stimulus-outcome associations

posterior/superior SFG ((Pre)SMA)

• > action selection
• response competition/comparison, conflict resolution (e.g. via automatic inhibition)
• Task 5: mOFC/vmPFC for object selection/comparison
• Task 1: (Pre)SMA, vPreMC for action selection (Wong)

Question 4

A3: Describe the experiment by Camille et al. (2011) on stimulus & action value in lesion patients (bridge between task 5 & 6)
- Methods

Answer 4

Subjects:

• lesion in mOFC/vmPFC (S-R)
• lesion in mCC (A-O, action selection)
• healthy control

2 reversal learning tasks:

• Stimulus value tasks
• > S-R associations are learned & relearned
• > choice between 2 decks of cards
• > each choice results in either win or loss (monetary feedback)
• Action selection task
• > A-O associations are learned & relearned
• > choice between 2 movements (internal/external twisting of wrist)

Both

• > one choice associated with better reward probability
• > after 13 correct trials, probabilities are reversed

Question 5

A3: Describe the experiment by Camille et al. (2011) on stimulus & action value in lesion patients (bridge between task 5 & 6)
- Results

Answer 5

mOFC/vmPFC lesion:

• selective impairment for stimulus-reward task (no deficit in action task)
• > more errors learning S-O
• > completed fewer reversals
• > more likely to switch choices after a win

mCC lesion:

• selective impairment for action-reward task (no deficit in stimulus task)
• > more errors learning A-O
• > more likely to switch choices after a win

Conclusions:

• OFC & MCC have dissociable roles in value-based decisions -> stimulus/action selection
• not fundamental for learning (more errors but still possible!)
• Evidence for Rushworth’s division

Question 6

A4: Thinking about the similarity of MCC & vmPFC/OFC (action/stimulus selection), why is it not surprise that the MCC is critical for outcome monitoring?

Answer 6

• vmPFC also involved in reward-based decision-making/reward expectation
• Reward-processing & performance monitoring have a close relationship
• role in determining the reward prediction error (striatum)
• > likely MCC has similar role for actions
• support by Camille study

Question 7

A4: Explain the Ullsperger et al (2003) experiment on brain systems involved in outcome monitoring/error detection, with MCC in central position.

Answer 7

• Participants observe short sequence of motion of 2 balls moving towards a finish line
• after 1.43s balls disappeared & subjects had to predict which ball would first cross the finish line
• Measured activity increases

1) error always followed by negative feedback & correct responses always followed by positive feedback
- -> informative on all trials

• Error + negative feedback: “I am uncertain + feedback is certain”
• > rCMA, Anterior Inferior Insula, Habenula, (Pre)SMA
  - > brain response to error & omission of reward
  - > could also represent response conflict & uncertainty
• Correct + positive feedback: “I am certain + feedback is certain”
• > vS/NA, putamen

2) 26.5% of trials contained no feedback information -> uncertainty

• Error + negative feedback: “I am uncertain + feedback is certain”
• > rCMA + insula
• Error + no feedback: “I am uncertain + feedback is uncertain”
• > no data
• Correct + positive feedback: “I am certain + feedback is certain”
• > vS/NA
• Correct + no feedback: “I am certain + feedback is uncertain”
• > habenular complex

Distinct activation for reward (vS) and no reward (rCMA + insula)
Distinct activation for unexpected non-reward
(- kinda like negative prediction error?)
- influenced DA neurons -> restrains midbrain nuclei -> influences reward prediction error calculation

Question 8

A5: Describe the phenomenon and the circumstances that are critical for Error-related negativity (ERN)

• ERN
• rCMA & habenula
• Error detection vs. response conflict monitoring

Answer 8

ERN

• ~60-120ms after error response
• Negative, upward deflection
• Fronto-central scalp signal

Ullsperger: ERN generated by MCC

• rCMA activates in response to errors if followed by negative feedback
• Habenula indirectly influenced rCMA by selectively increasing its activity for negative feedback on errors
• > ERN reflects error detection
  - > independent of whether subject is conscious of error

Rushworth: Dissociation of error detection & response monitoring

• MCC sulcus/rCMA -> error detection
  
  • > mediates cost-benefit decisions by comparing expected outcome with cost of action
  • > Reward-guided action selection + learning of action-outcome associations
• TMS lesion -> reduced ERN, intact error awareness
• SFG ((Pre)SMA), SEF) -> action comparison, response conflict monitoring
  
  • > task control + selection of action sets
  • > necessary when action selection rules are first implemented & changed (set-switching, conflict)
• TMS lesion -> less sensitivity to response conflict

Question 9

A6: How does error monitoring show up in forms of psychopathology

• OCD
• Depression
• Schizophrenia

Answer 9

Hyperactive error-processing signals in OCD and depression

OCD

• Larger ERN
  
  • > ERN magnutide correlates with OCD symptoms
  • > Possible vulnerability factor for OCD
• Excessive task-related activity in MCC
  
  • > increased resting activity in MCC
  • > Explanation: problems with performance monitoring in OCD

Depression

• Larger ERN & feedback negativity
  - > Negative affect (both in depression & OCD) may be associated with higher ERNs
  - > some studies: no difference/reduced ERNs
• Association between ERN magnitude & S-allele of 5-HTTLPR
  
  • > associated with neuroticism &life stress

Schizophrenia

• Smaller ERN
• Diminished MFC signals in response to errors (error detection & response conflict)
  - > rACC is source of this decrease (no dACC)
  - > interpreted as reflecting motivational deficit (negative schizophrenic symptom) interacting with performance monitoring
  - > Inverse correlation between ERN magnitude & negative symptom severity

Question 10

Theoretical relation of ERN & emotional responses

Answer 10

Error = negative emotional response

• may be important part of measured scalp potential
• rACC might process emotional component of error signal
• > has more projections to limbic system (e.g. AMY) than cACC
• MCC: cognitive acativity
• rACC: emotional activity

Response interference task

• errors -> affective response to a stimulus would reflect the salience & value that an individual attaches to a stimulus
• errors leading to monetary loss -> greater rACC activity vs. errors where no money is lost (or gained)
• OCD
• > greater value on performing correctly -> may drive excessive ERN in rACC