Learning Pt1 PBS5

The acquisition of a new behavioural (or physiological) response to a previously neutral stimulus as a result of experiencing a predictive relationship between  it and a biologcally relevant stimulus.

• Unconditioned: a stimulus that has natural relevance - behaviour not conditional on the learning experience (eg. food)
• Conditioned: a stimulus that gains its relevance through learning - behaviour conditional on learning experience (eg. bell)
• Conditioned Response (CR): response to the CS after training - ability of CS to elicit this response is conditional on the learning experience.

• 1. Appetitive (association with pleasant US)
• 2. Aversive (association with unpleasant US)
• Increase in vigour of CR with strengthening of association between US and CS. 

• Role of dopamine (DA) system
• Cells in substantia nigra and ventral tagmental area (VTA) in brain stem project to...
• many forebrain structures, particularly striatum,
• where they facilitate neural processing by releasing dopamine as neurotransmitter.

• O'Doherty et al (2002)
• Conditoning either to appetitive glucose solution following one picture CS and aversive salt solution to another CS
• fMRI during presentation of 2 CS
• Substantia nigra and VTA more active during CS associated with appetitive glucose US.

• dopamine in addiction
• from dorsal striatum --> ventral striatum (check)

• 1. change in behavioural (physiological) response to the CS
• 2. change in cognitive expectancy of US following presentation of CS
• Often, the two are in concordance. Is there a causal relationship between cognitive expectation and acquisition of behavioural response?

• Lovibond (1992)
• Paired one pic of plant with electric shock (CS+), another with no shock
• Intermixed
• Participants rate likelihood of shock to each CS and their skin conductance response (SCR) also measured
• Those who rapidly learnt this discrimination (with correct predictions) also showed SCR on CS+ trials
• No SCR in participants who were unaware of relationship between CS and US

• Hugdahl & Ohman 
• Instructed participants there would be no more shock in an extinction phase (repeated presentation of CS without US)
• Usually, extinction typically declines steadily across trials
• BUT: instructed extinction resulted in immediate disappearance of SCR
• This is explicit learning - where awareness of relationship between CS and US causes presence or absence of behavioural response

• Hugdahl & Ohman (again)
• if using a fear relevant stimuli (instead of plants) like spider or snakes, then even when participants are told that these stimuli will no longer lead to shock had little impact on extinction. 
• Biological preparedness 
• Implicit learning seems independent of cognitive expectation and occurs when CS signals a potentially dangerous aversive US

• Bechara et al (1995)
• 3 patients with bilateral damage to subcortical structures in temporal lobal
• amygdala (AMG), hippocampus (HC), amygdala+hippocampus (AMG+HC)
• AMG and AMG+HC patients failed to acquire SCR response to CS
• But HC and HC+AMG damage patients couldn't report accurately when asked which CS was folowed by the US, whereas AMG patients could.
• Double dissociation between implicit SCR learning mediated by AMG, and explicit cognitive learning mediated by HC

• The time between CS and US
• Important factor in establishment of CS-US relationship
• Rate at which rats press lever for food decrease with delay between the press and food delivery
• BUT: high temporal contiguity not necessary for conditioning --> depends on kind of learning involved

• Conditioned food aversion - learning occurs with long intervals between taste and later sickness
• Andrykowski & Otis (1990)
• Interview patients about type of food consumed prior to chemotherapy (many receiving chemo often develop aversions to food)
• Food eaten closer to chemotherapy were not more likely to develop aversions
• No relationship between time a food was eaten (in relation to chemo and relation to vomiting) and whether patients developed aversion to it

• When a stimulus A is paired with US first, and then A and B are paired together with US, then when stimulus B is presented alone, one has not learned the CR associated with US. 
• Process:
• Stage1: A associated with US. A leads to CR.
• Stage2: AB is associated with US. B alone does not lead to CR. B and US is not associated.
• Control: BC associated with US (C is new stimulus). B leads to CR.

• Resulting from predictive learning
• A predicts outcome
• In presence of A, B is not leared about outcome, because A allows US to be expected
• Learning occurs when there is a discrepancy between how much US is expected given a CS.

• Tobler et al (2006)
• Ventral striatum significantly lower activation to blocked reward-predicting stimuli rather than non-blocked
• In reward-predicting control situations OFC and ventral striatum deactivated when reward left out
• Responses in discrete parts of OFC correlated with degree of behavioural learning during and after the learning phase.
• CHECK

• unexpected 
• negatively accelerated curve
• expectation 
• increases
• decreases

• Rescorla-Wagner rule
• ∆V_A = αβ(λ – ∑V_{all CSs})
• V= associative strength of CS (expectation)
• ΣV(the last one) = sum of current associative strength of all available CS on that trial
• αβ= learning rate (α=saliency of stimulus, β=strength of reinforcer)
• λ= perfect prediction of US
• (λ-V)= prediction error (surprise)

• Increase in associative strength of CS result from degree to which current associative strength deviates from perfect learning. This deviation is the prediction error. 
• When US is present, λ=1 , and associative strength is excitatory (anticipation of presence of US)

• ΣV is already at 1
• λ=1 because US is present
• Which means λ-ΣV = 0
• Which means Change in V = 0.

• 1. A-->US
• AB--> US
• B -->No US
• 2. A--> US
• AB--> No US
• CB --> learning to C is much slower with B which shows that B is an inhibitor