Difficulty: Intermediate | Prerequisites: Parts 1 and 2 of these notes. Familiarity with prefrontal cortex anatomy and the basal ganglia is helpful but not essential.
Parts 1 and 2 covered what emotions are and which brain structures underpin specific emotions. This final section shifts to motivation: how the brain computes whether something is currently worth pursuing, evaluates the costs and benefits of an action, and predicts future rewards. Three regions take centre stage: the orbitofrontal cortex (OFC) for tracking current reward value and enabling behavioural flexibility, the anterior cingulate cortex (ACC) for evaluating actions and effort, and the ventral striatum for reward prediction via dopamine signalling. Together they explain why you sometimes resist a piece of chocolate even though you like it, and why unexpected rewards feel better than expected ones.
The OFC computes the current reward value of stimuli, giving you the flexibility to stop eating chocolate when you are full. The ACC evaluates the value of actions, weighing effort against reward. The ventral striatum (including the nucleus accumbens) signals reward prediction errors via dopamine: unexpected rewards produce a burst, expected rewards produce nothing, and omitted expected rewards produce a dip. OFC damage leads to impulsivity and poor social regulation; ACC damage leads to "laziness" (choosing least effort). These three regions work together as a circuit for motivation and reward-based decision-making.
Orbitofrontal cortex (OFC)
A region of the prefrontal cortex located above the eye sockets. It computes and updates the current reward value of stimuli, enabling behavioural flexibility. Think of it as the part of the brain that decides whether something is worth having right now, given context.
Anterior cingulate cortex (ACC)
A region wrapping around the corpus callosum with two functionally distinct subdivisions: a dorsal ("cognitive") division and a ventral ("affective") division. The ACC evaluates the value of actions and monitors effort-reward trade-offs.
Ventral striatum
Part of the basal ganglia, including the nucleus accumbens. It receives dopamine inputs from the midbrain (ventral tegmental area) and signals reward prediction. Identified as a "reward centre" through self-stimulation studies.
Nucleus accumbens
A key structure within the ventral striatum that receives dopamine projections from the midbrain. Implicated in the hedonic aspects of reward, including drug use.
Reward prediction error
The difference between the reward you expected and the reward you received. Dopamine neurons fire more when a reward is unexpected (positive prediction error), do not change when a reward is expected and delivered, and reduce firing when an expected reward is omitted (negative prediction error). Described by Schultz et al. (1997).
Maslow's hierarchy of needs
A motivational framework (1943) proposing that human needs are arranged hierarchically: physiological needs at the base, then safety, love/belonging, esteem, and self-actualisation at the top. The course emphasises social needs but notes the hierarchy is not necessarily rigid.
Behavioural flexibility
The ability to adjust behaviour when the reward value of a stimulus changes. OFC damage compromises this: patients continue to pursue previously rewarded stimuli even after the reward is removed.
Impulsivity (OFC-related)
A tendency to choose small immediate rewards over larger delayed ones, or to take excessive risks. Seen after OFC damage in both rats and humans.
Maslow proposed five levels of motivation arranged as a hierarchy: physiological needs, safety, love/belonging, esteem, and self-actualisation. The model usefully emphasises social needs alongside basic survival needs. However, evidence suggests needs are not strictly hierarchical in practice.
Deciding whether to eat a piece of chocolate is not just about remembering that chocolate is pleasant. You also need to evaluate the current context: are you already full? Does the chocolate belong to someone else? The OFC computes whether a stimulus is currently rewarding.
Key studies:
Small et al. (2001): PET study of eating chocolate. Initially the chocolate was wanted and pleasant, but after excessive consumption it became unwanted and unpleasant. OFC activity tracked this transition, shifting from medial to lateral activation.
Plassman et al. (2008): fMRI of wine tasting. Wines believed to be more expensive tasted subjectively nicer, and these subjective ratings correlated with OFC activity. The OFC tracks perceived, not just actual, reward value.
Blood and Zatorre (2001): Subjective pleasantness of music correlated with OFC activity, showing the OFC codes reward value across sensory modalities (not just taste).
Computing current reward value gives behavioural flexibility: you can change course when something stops being rewarding. Patients with OFC lesions lose this flexibility and behave impulsively:
Rudebeck et al. (2006): Rats with OFC lesions prefer small immediate rewards over larger rewards offered later.
Human patients fail to regulate behaviour according to social norms, show inappropriate emotional responses, and have poor social relationships.
Rolls et al. (2004): Patients fail to shift away from a previously rewarded stimulus that is no longer rewarded (perseveration).
Bechara et al. (1994): Patients take higher-risk gambles, consistent with an inability to update the expected value of choices.
The ACC has two functionally distinct parts:
Dorsal ACC ("cognitive"): involved in conflict monitoring, error detection, and evaluating the value of actions. Activated during tasks like the Stroop test, where conflicting information must be resolved.
Ventral ACC ("affective"): involved in emotional processing. Activated by tasks involving social comparison (e.g. evaluating one's position relative to others; Gianaros et al., 2007).
Rushworth et al. (2007) proposed a clean dissociation:
The dorsal ACC computes the current value of actions (is this action worth the effort?).
The OFC computes the current value of stimuli (is this thing currently rewarding?).
This was supported by lesion studies in rats (Rudebeck et al., 2006):
Rats with ACC lesions become "lazy," favouring the option requiring least effort to obtain a reward.
Rats with OFC lesions become "impulsive," favouring whichever option delivers a reward soonest.
Blair et al. (2006) found greater dorsal ACC response when comparing similar-magnitude losses ($50 vs. $100) than when comparing very different losses ($50 vs. $300). This is consistent with the ACC evaluating difficulty of choice (closer options are harder to decide between).
Part of the basal ganglia. Forms a loop with the OFC and thalamus.
Includes the nucleus accumbens, which receives dopamine inputs from the midbrain (ventral tegmental area and substantia nigra).
Identified as a "reward centre" through classic self-stimulation studies (Olds, 1956): rats with electrodes in this region would press a lever repeatedly to stimulate their own brain, sometimes to the exclusion of food and water.
Increased dopamine release occurs when animals are presented with secondary reinforcers paired with food, or with social rewards (e.g. male rats presented with a female; monkey "pay-per-view" studies).
In humans, the ventral striatum is implicated in the hedonic aspects of drug use (e.g. cocaine releases dopamine here).
In human fMRI, activity in the ventral striatum correlates with the level of potential monetary reward.
Dopamine neurons do not simply signal that a reward has arrived. They signal the difference between expected and actual reward:
No prediction, reward occurs: dopamine neurons fire at the time of reward delivery. The reward was unexpected, so there is a positive prediction error.
Reward predicted, reward occurs: dopamine neurons fire at the time of the predictive cue (CS), not at the time of reward delivery. The reward was expected, so there is no prediction error at the time of delivery.
Reward predicted, no reward occurs: dopamine neuron firing drops below baseline at the time the reward was expected. The reward was omitted, producing a negative prediction error.
This pattern means dopamine encodes the surprise value of rewards, not reward itself. It is fundamental to reinforcement learning: the brain updates its predictions based on whether reality matched expectations.
Students often think the OFC and ACC do the same thing. They do not. The OFC evaluates the current value of stimuli; the ACC evaluates the current value of actions and effort.
Dopamine is frequently described as the "pleasure chemical." Schultz's work shows dopamine neurons encode prediction error (surprise), not pleasure per se. A fully expected reward produces no dopamine burst at the time of delivery.
Maslow's hierarchy is sometimes treated as neuroscientific fact. It is a motivational framework, not a neural model. The course uses it to illustrate that social needs are important, not to endorse a strict hierarchy.
OFC damage is sometimes confused with general frontal-lobe impairment. OFC lesions produce a specific pattern: impulsivity, failure to update reward values, poor social regulation. They do not produce general cognitive impairment.
The OFC-ACC distinction (stimuli vs. actions) is a high-value exam topic. Know the Rushworth et al. (2007) framework and the Rudebeck et al. (2006) rat lesion findings that support it.
Be able to describe at least two studies showing OFC activity tracks current reward value (Small et al., Plassman et al., or Blood & Zatorre).
Know the three patterns of dopamine neuron firing in Schultz et al. (1997) and be able to explain reward prediction error in your own words.
OFC lesion effects are commonly tested: impulsivity, perseveration, higher-risk gambling, poor social regulation. Know at least two studies.
Understand how the ventral striatum, OFC, and thalamus form a reward circuit.
True or False: The OFC computes the current value of actions. (False: the OFC computes the current value of stimuli; the ACC computes the current value of actions.)
Fill in the blank: In Schultz's reward prediction error model, dopamine neurons fire below baseline when a predicted reward is ______. (Omitted / not delivered.)
True or False: Rats with ACC lesions are impulsive, choosing the fastest reward. (False: ACC-lesioned rats are "lazy," choosing least effort. OFC-lesioned rats are impulsive.)
Fill in the blank: The nucleus accumbens is a key structure within the ______ and receives dopamine from the midbrain. (Ventral striatum.)
True or False: OFC activity tracks only the actual reward value of a stimulus, not its perceived value. (False: Plassman et al. showed OFC activity tracks perceived value, as wines believed to be expensive were rated nicer and correlated with OFC activity.)
Q: Explain how the OFC tracks current reward value, using at least two studies as evidence.
A: The OFC computes whether a stimulus is currently rewarding by integrating its emotional history with the present context. Small et al. (2001) showed this with a PET study: chocolate was initially pleasant (medial OFC active), but after excessive consumption it became unpleasant (lateral OFC active), and OFC activity tracked the transition. Plassman et al. (2008) demonstrated that wines believed to be more expensive tasted subjectively nicer and correlated with OFC activity, showing the OFC codes perceived value, not just actual sensory input.
Q: What is the key functional distinction between the OFC and the dorsal ACC, according to Rushworth et al. (2007)?
A: The OFC computes the current value of stimuli (is this thing rewarding right now?), while the dorsal ACC computes the current value of actions (is this action worth the effort?). This was supported by Rudebeck et al. (2006): OFC-lesioned rats became impulsive (choosing fastest reward), while ACC-lesioned rats became lazy (choosing least effort for a reward).
Q: Describe the three patterns of dopamine neuron firing in Schultz et al.'s (1997) reward prediction error model.
A: (1) When no prediction is made and a reward occurs, dopamine neurons fire at the time of reward delivery (positive prediction error: the reward was a surprise). (2) When a reward is predicted and occurs, dopamine neurons fire at the time of the predictive cue but not at reward delivery (no prediction error: the reward was expected). (3) When a reward is predicted but does not occur, dopamine neuron firing drops below baseline at the expected time of delivery (negative prediction error: the omission was a surprise). This means dopamine encodes the surprise value of rewards, not reward itself.
Q: What behavioural effects result from OFC damage?
A: OFC damage leads to impulsivity (preferring small immediate rewards over larger delayed ones), failure to update behaviour when reward contingencies change (perseveration, as in Rolls et al., 2004), higher-risk gambling (Bechara et al., 1994), and difficulty regulating behaviour according to social norms, leading to inappropriate emotional responses and poor social relationships.
The OFC's role in computing reward value connects to later material on social decision-making and moral judgement, where people must weigh immediate selfish rewards against social norms. The ACC's role in effort evaluation links to research on empathy for pain (anterior cingulate activation when witnessing another's pain). Reward prediction error and dopamine signalling are foundational for understanding addiction, habit formation, and reinforcement learning models used across cognitive neuroscience. The ventral striatum's involvement in social reward (Rilling et al., 2002, from Part 1) ties the motivation circuit back to the social dimension of emotions.
Orbitofrontal cortex OFC, anterior cingulate cortex ACC, dorsal ACC, ventral ACC, ventral striatum, nucleus accumbens, dopamine, reward prediction error, Schultz, reward centre, self-stimulation, Olds, Maslow hierarchy of needs, behavioural flexibility, impulsivity, perseveration, current reward value, Small chocolate PET, Plassman wine fMRI, Blood Zatorre music, Rudebeck lesion rats, Rolls reversal learning, Bechara gambling, Rushworth action value stimulus value, Blair loss comparison, basal ganglia, ventral tegmental area, substantia nigra, reinforcement learning, UA35, social neuroscience, NYU chapter 4 motivation reward punishment