Lecture 15 (Executive Function/Cog. Control) Flashcards
What is executive function/cognitive control?
Ability to orchestrate millions of neurons to produce behaviour that is WILLFUL and COORDINATED.
Central role of PFC (connected DIRECTLY to every distinct functional unit of the brain)
Top-down vs bottom-up processing
Classical view was “bottom-up”/stimulus-response (behaviourist view). BUT this strategy is inflexible and potentially maladaptive.
“Top-down” –> brain must be able to control action to accomplish goals and select only those inputs that convey meaning for the desired action
Concepts of bottom up and top down can be ambiguous (see slides)
Stephen Grossberg
One of the first supporters of the “active” brain. In his model, feedback has a role in focusing selective attention to important elements of the environment.
Actual vs expected input is compared. Feedforward: adaptive filtering of inputs. Feedback: predictive signals for expected patterns.
PFC
The “great integrator” - remote and local inputs, most connections = reciprocal, access to internal and external information
Sensory, motor and limbic regions
SWRs
Studies have shown when rat runs along linear track, place cells fire in particular place (place field).
Sharp wave ripples (SWRs) = fast oscillations, seen “riding” on sharp waves in hippocampus. Due to firing of APs by groups of pyramidal cells.
Ripples can predict order of movements (e.g. route rat will take). Provides info to PFC about potential route. Also provides info to PFC about what movement was made (played in reverse order after movement)
Also gets replayed during sleep (–> LTM?). Only during slow wave sleep (also played faster). CA1 cells of hippocampus are paired with PFC cells, which fire 100ms of one another. This does not occur during REM.
Phineas Gage
Major medial PFC damage –> major personality change. Previously efficient worked –> profane, rude, unable to organise work/life. Loss of cog. control.
Ability to perform complex tasks is intact but lack ability to form sensible long-term concerns and goals.
Dysexecutive syndrome.
Tests of PFC damage
TOWER OF LONDON TASK: tubes with balls, given a start state and goal state. Need to use as few moves as possible. PFC damage –> many more moves used due to deficit in forward planning.
WISCONSIN CARD SORT TASK: cards have different shapes on them, in different numbers with different colours. Need to learn matching rule (e.g. same colour, same number etc) by putting cards together and being told if right or wrong. PFC damage –> can learn rule but can’t learn new rule if it’s then changed.
STROOP TASK: word for colour written in different colour. Asked to name the colour in which the word is written. Inability to inhibit automatic responses.
IOWA GAMBLING TASK: given 4 decks of cards. Each turn, pick a deck. Each turn, get a win or loss. Decks A and B have larger wins, but also more frequent and larger losses. Optimal long-term strategy is therefore to stick with C and D.
Control: people tend to pick randomly, then pick A and B, but then work it out and switch to C and D.
PFC damage: tend to stick to A and B (can’t switch to optimal strategy)
In normal subjects, get activation of PFC during task, along with memory system and emotional system (amygdala)
Also measure SCR (skin conductance response - emotional response –> more sweat/lower resistance). Normal subjects: before choosing, larger response for A & B than C & D (because more risk).
Amygdala lesions: response much lower, no difference between the letters.
PFC lesions: slightly larger response than amygdala lesions (because still have amygdala intact) but no difference between letters.
PFC lesions in marmosets (2 experiments)
Given option of two shapes with lines through them. First have to select correct one based on shape alone.
Trained to criterion of 1 of 3 possible rule changes:
1) intradimensional shift (new shape gets reward)
2) extradimensional shift (new line gets reward)
3) reversal (opposite line gets reward)
Lateral PFC lesion inhibits (2). Orbital PFC lesion inhibits (3). Neither affect (1).
Deficit is specific to inhibiting a PREVIOUSLY LEARNED response (both lesions can learn rule, but can’t change to new rule)
Other experiment: sample put in one of two holes, delay, must choose correct one (SPATIAL DELAYED RESPONSE) - sensitive to dorsolateral PFC lesions.
NON-SPATIAL (DMTS): two samples cover holes, have to pick correct sample - sensitive to ventrolateral PFC lesions.
Reversal of either of these (spatial or non-spatial) tasks - sensitive to orbitoventral PFC lesions
Suggests dorsal –> spatial memory
Ventral –> object memory
Orbital –> inhibitory control
HOWEVER, dorsolateral lesions do produce deficit in non-spatial task e.g. 3 objects. Trial 1) pick random object. Trial 2) must pick different object to trial 1. Trial 3) must pick final object. This is a non-spatial task (doesn’t matter where objects are) BUT contains both spatial and object components. Unlike previous tasks (above), they aren’t segregated. Lesion to either will –> deficit.
PFC cells in delay tasks
Many PFC cells maintain information about spatial or object cues during delay periods. Can encode all aspects of a trial.
Animal fixates, cue appears. Delay. Then has to move focus to position of cue.
Some units respond during the delay, but only for a particular upcoming movement (to particular location)
Dorsolateral PFC neurons show memory for visual field location. Resistant to visual distractors in preferred location (unlike sensory cortex neurons). Responses may depend on signal that predicts reward (dopaminergic) - dopamine antagonists block delay activity.
PFC cells respond to RULE rather than particular cue/object
E.g. 3 tasks: spatial (animal must fixate on correct space), object (must fixate on correct object), associative (animal must infer from object which direction to fixate in).
In object vs associative task are getting given the same cue, but neuron only responds if the task is associative.
PFC neurons encode CATEGORIES
6 images, ranging from 100% cat to 100% dog (morphs)
Can respond correctly/discriminate all of them, even 60% dog or 60% cat.
Irrespective of percentage, firing of particular neuron is the same for all 3 dog images. Different response for all 3 cat images.
Cells are operating in rule-based fashion (not simply responding to stimulus)