Dopamine 1 & 2 Flashcards
what is the evidence for divergent signalling of DA neurons
Matsuda et al., 2015
traced axonal arborisations of nigrostriatal dopaminergic neurons that were viral vector expressing of green fluorescent protein
found highly dense arborisations that were widely spread in the rat brain (likely scales to 1-2.5 million in humans)
one to many system of divergent signalling is key to dopamine neurons’ ability to broadcast a signal that exerts a strong influence over many striatal neurons which is key to function in dopamine based learning
also indicates that neurodegeneration of a single nigral neuron can affect multiple neurons in the striatum
Describe the neuroanatomy of midbrain dopamine
Ventral tegmental area (VTA) - mesolimbic pathway
Projects to nucleus accumbens and dorsolateral striatum
Substantia nigra pars compacta (SNc) - nigrostriatal pathway
Projects to dorsolateral and dorsomedial striatum
what is the evidence from early animal studies that dopamine reinforces choices/behaviour
(Olds & Milner, 1954)
Tested the effect of electrical stimulation on operant response learning
Inserted electrodes into different areas of a rat brain and recorded rate of self-stimulation (lever pressing)
Identified hotspots around the lateral hypothalamus and in the midbrain - corresponding to the VTA associated with high levels of lever pressing
Stimulation of VTA reinforces behaviour of pressing not seen for activation in other areas
what is the evidence that dopamine only signals RPE, rather than updating choice policy
Blanco-Pozo et al., 2021
Mice were trained on a behavioural task where they were given either a free choice between left and right illuminated nose pokes or forced choice. After the first choice was made a cue would sound to indicate whether the mouse should go to the upper or lower cubby to receive a reward
the probability that reward would be delivered in the upper/lower cubby changed with blocks but the probability of a left/right nose poke leading to the auditory cue for the upper/lower cubby was fixed. Thus the value of the first action was fully predicted by the latent variable of reward probability.
Recording from VTA, NA, DMS using fibre photometry to record Ca2+ signals indicated that the dopamine signals for the first and second choices were strongly influenced by value and reflected reward prediction error. Recent reward rate explained DA activity on a slower time scale
But this is correlational evidence and does not prove that the signal is driving animals’ choice behaviour.
Mice expressing ChR2 or control in VTA DA neurons and examined the effect of optogenetic stimulation after the first-step choice (time of second-step cue) or at outcome cue presentation
Stimulating dopamine after first-step choice was reinforcing and significantly increased the probability of repeating that choice
Stimulating dopamine at the time of trial outcome had no subsequent effect on choice
Reinforced immediately preceding actions but failed to recapitulate behavioural consequences of natural rewards at outcome despite rewards driving the largest dopamine signals in the task
When inference guides choice, rewards have a dopamine-independent influence on policy through the information they carry about the world’s state
what is the early evidence from animal models that dopamine mediates the rewarding effects of hedonic value
Wise et al., 1978
Mice treated with pimozide, a D2 receptor antagonist
Lever-pressing and running for food reward attenuated in hungry rats
Ruled-out drug-induced performance difficulties due to periods of normal responding
Selective blunting of the rewarding impact of food and other hedonic stimuli
what is the early evidence for DA spike activity reflecting a prediction error
Schultz, 1997
Dopamine activity precisely correlates with RPE in primates
Almost dopamine activity in the lateral VTA is consistent with RPE coding
Before learning a drop of appetitive fruit juice occurs in the absence of prediction
This produces a positive error in the prediction of reward, producing a DA spike
After learning the CS (light) predicts the stimulus, DA spike for the CS but not the reward
After learning the CS predicts the stimulus but no reward occurs due to a mistake in the behavioural response, DA spike for CS but depression for absence of reward at the precise time the reward would have occurred
what is the causal evidence for DA RPE driving cue-learning (and limitations to this research)
Steinberg et al., 2013
Measured the effect of optogenetic stimulation of VTA neurons on pavlovian blocking and extinction
Rats were trained to go to reward port for sucrose reward (US) after hearing an auditory cue (CS)
as the occurrence of the sucrose delivery is fully predicted by the auditory cue, no new learning occurs when the auditory cue is paired with the light cue, as tested by behavioural response to light cue a day later.
however transgenic cre recombinase expressing rats injected with cre dependent channelrhodopsin2 into VTA neurons activated by light recieved optogenetic stimulation either at the delivery of the US or during the ITI
delivery of stimulation at US resulted in learning about the light cue (as tested 24hrs later) reflecting behaviour of control group who were pre-trained with a different auditory cue
same seen in extinction paradigm where sucrose reward is removed or replaced with water
BUT extinction seems to occur on at time t rather than t+1 so behaviour changes before / alongside prediction error
also duration of stimulation far exceeds natural DA spike
also blocking is a pavlovian phenomenon but this design is instrumental
