Cognitive Control Flashcards
Reacting fast
When actions are very fast, the system is advised to use the most evident response to take a decision (direct pathway)
Quand l’ation va très vite, le sytème utilise automatiquement la réponse la plus évidente pour prendre une décision : voie directe.
Executive functions ?
Flexibility, planning, inhibition, switch, attention, monitoring High level processes that control a planned action
When cognitive control should be mobilized ?
> In law predictible env, where changes are frequent
> When the risk of error is high (little time or too many options)
> When consequences can be dramatic
[Resisting to the call of the Sirens]
Ridderinkhof
How to study control processes ?
To do a relatively simple task but in a sub-optimally manny
example: we can use temporal pressure (errors will be done because of the temporal pressure = as fast as possible)
This is the Reaction Time task (RT task)
Speed-accuracy tradeoff
Courbe d’échange vitesse-précision
The relationship between the mean RT and the precision of the response follows a curve called Speed-accuracy tradeoff (courbe d’échange vitesse-précision)
Theorical RT / Empirical RT
Stimulus-Response Compatibility
The stimulus-response compatibility is the degree to which a person’s perception of the world is compatible with the required action.
La compatibilité stimulus-réponse est le degré auquel la perception d’une personne est compatible avec la tâche requise.
S-R compatibility has been described as the « naturalness » of the association between a stimulus and its response, such as a left-oriented stimulus requiring a response from tje left side of the body.
Spatial / symbolic
La compatibilité S-R est une de naturalité de l’association entre le stimulus et sa réponse comme par exemple un stimulus présenté à gauche pour une réaction devant survenir à gauche.
SR Compatibility : cognitives ergonomics
Behavioral studies in conflict tasks suggest the existence of an online monitoring, i.e. during the course of a trial.
RT are shorter when stimuli are compatible compared to incompatible
The error rate is higher in incompatible situations.
Les études comportementales de tâches conflictuelles suggèrent la présence d’un monitioring online, durant les essais.
Les TR sont plus courts pour des stimuli compatibles comparés aux stimuli incompatibles.
Les taux d’erreurs sont supérieux pour les situations incompatibles.
SRC : others tasks
Simon Task, Eriksen Flanker task (target is the letter on the center, and the two flankers on the both sides have to be ignored), Stroop Task
Dynamic of the conflict effet
The shape of the RT’s distribution is interesting to study cognitive control mechanisms.
Probability Density Function : probability that the response occurs for a certain value of the post-stimulus time.
Cumulative Probability Function: probability that the response has been given at a certain latency.
Dual route model
Dual Route model
The stimulus’s position activate automatically the ipsilateral response
The processing of the color activates in a controlled way the response associated to the instructions
The controlled processing are slower than the automatically activation.
At the responses activation level, a inhibitory process acts on the response representation that has to be contained.
(ex de Simon task)
Delta curves
Delta curves show the convergence in the compatible and incompatible distribution
—> the compatibility effect decrease when RT increase
—> this would be the reflect of an active inhibition of the automatic activation of the ipsilateral response to the stimulus
For short RTs—> inhibition did not have the time to build and the correct response is delayed (in time) by the activation of the incorrect response
For long RT’s —> the incorrect response is more inhibited.
Delta curves
What suggests an online control?
DELTA CURVES
- *For short RTs →**
- The inhibition did not have the time to build
- The correct response is delayed by the activation of the incorrect response
Si TR long →
The incorrect response is more inhibited
The effect due to position of the stimulus is reduced
The difference between compatible and incompatible trials is weaker
Work of Richard Ridderinkhof (Amsterdam University)
Cf in bibliography Ridderinkhof (AttPerf, 2002)
Errors
Interesting to study coz frequent in daily life and useful for learning
→ Errors are more frequent in incompatible situations
→ Error RTs are faster than correct RTs. Why ?
CAF : Conditional Accuracy Function
An important slope in the CAFs reflects a higher proportion of fast errors. This suggests a higher initial activation of the incorrect response (Ridderinkhof, 2002; Wylie et al., 2009).
reactive VS proactive control
Action control aims
- To select the relevant actions for the current situation
- To resist to the tend to respond quickly to non relevant responses.
- To anticipate the difficulties
• Online action control or Reactive control : refers to the processes acting to suppress online an incorrect, not appropriate or not desired action.
requested when subjects are in an automatic mode
REACTIVE > ACC
• Anticipatory action control or Proactive control : refers to processes which helps to avoid making an error and to be disturbed by irrelevant information.
involved when subjects are in a controlled mode
PROACTIVE > DLPFC
Behavioral adjustments
Post-Error Slowing
reducing temporal pressure to gather more evidence before selecting the response. The error has to be perceived (consciously or not)
These post-error adjustments which are very important in trial-and-error learning processes, cannot inform on the detection mechanisms of the errors, and neither on the capacity to inhibit the errors detected on time.
Compatibility or congruency effect
Longer RT’s for incompatible trials
Sequential effect : Interference effect
The interference effect is observed after an error.
(Reduction of the compatibility effect after an error)
Sequential effect : Gratton effect
After a compatible trial, large compatibility effect
After an incompatible trial, small even absence of the compatibility effect
Proactive mechanisms of control are trigger after error but also after a risky situation.
Pure correct trials
Pure-error trials
Incorrect - Correct trials
Pure-Correct trials
RT = PMT + MT
Pure-Error trials
RTe = PMTe + MTe
Incorrect - correct trials
RT = PMT + MT
PMT = IA + CT
Incorrect activation + Correction time
Presence of a Partial Error
Partial Errors
Partial errors were :
- Detected
- Inhibited
- Corrected 
In these trials, on line control mechanisms do occur ! Burle et al. (Psychological Research, 2002)
Full error : control mecanism
Less force is used to press wrong types
EMG enables to better understand this effect :
→ Chronometric index
RTc > RTe : RTs are shorted in errors
PMTc > PMTe : PMTs are shorted in errors
MTe > MTc (le temps moteur des erreurs est plus long <= online control)
→ Physiological index
The system detected the error and attempted to stop it
EEG
NE or ERN (70-100ms)
(Falkenstein’s group)
source in the rostral cingulate zone ACC and/or SMA
> Error and incompatible trials
> Carter 98 : Response competition area (formerly error detection mechanism)
Partial error NE ?
The size of the amplitude depends on how strong is the partial error. (PE>IC>C=0)
This can be very interesting for brain computer interface : use the brain to pilote a chair for handicapped people. To know when the person wants to stop.
PE (200-300ms)
conscience de l’erreur
Other EEG Activities linked with control processes
VanVeen & Carter (2002)
N2 in correct trials < for incompatible
but not distinction btw pure errors & incorrect-correct trials (we know that elicit a N2)
Pe
only conscious errors are followed by post error slowing
Not observed after PE
- - > consciousness of error ?
The negativity linked to feedback : FRN
“Feedback Related Negativity” (Miltner et al., 1997).
FRN and Ne seem to reflect similar physiological processes:
- External feedback for the FRN
- Internal feedback for the Ne
Observe someone doing error
Eriksen Flanker task
ERN & oERN seem to be the same
ERN amplitude = importance of making an error ?
Normal development
Children less perform but however fully conscious => post errors adjustements
Why ?
ACC imaturity
Emotionaly less implied in making errors than adults
Hypofunctioning
ERN amp smaller
Autism (immaturity of ACC?)
SCZ (ACC, DLPFC failure ?)
Hyperfunctioning
ERN amp bigger
OCD / Anxious disorder /
(ACC hyperactivated? more symptoms, more amplitude)
Alcohol - Fatigue - Religion
Alcohol
impair detection of performance/error
Decrease in interference effect
(diff btw congruent & incongruent larger after correct trial)
> no adjustement of behavior
Fatigue
capacity to detect errors OK but difficulties to adjust behavior
Fatigue + alcohol = real danger !
Religious
feel less responsible for their error
Cerebral substrate
Dopamine involved
> regulating ACC ?
ACC : 24-32 BA
SMA : 6-8
PROACTIVE > DLPFC
REACTIVE > ACC
basal ganglia : motor control
Conflict loop theory & model
Botvinick et al (2001)
model that explains adjustments in control
Outstanding questions
- None of the models can explain the Ne/ERN in correct trials.
This activity is a problem for the modelers. We should try to build a model taking into account this information. - None of the models can explain the ability to adjust online the behavior (cf partial errors and in full errors). They only explain the adjustments in control and not online control (proactive versus reactive control). Efforts have to be done also in this direction.
