The feedback-related negativity (FRN) is a negative event-related potential (ERP) component associated with the presentation of task feedback. The possibility that the FRN may habituate has been briefly mentioned in previous research (Garrido Chaves et al., 2020), but not yet been actively investigated. Thus, the current study is the first one to explicity investigate the possibility of short-term (across trials) and long-term (across blocks) habituation effects on the FRN. This was done by using electroencephalography (EEG) and a time-estimation paradigm during which participants were tasked with guessing the duration of one second. Following each estimate, participants were presented with either positive or negative visual feedback (however, only trials with negative feedback were included in the subsequent statistical analysis). It was hypothesized that mean FRN amplitude would decrease, i.e. habituate, upon increased exposure to negative feedback. Contrary to the expected effect, repeated-measures analyses of variance (ANOVA) and t-tests revealed a stong, significant sensitization effect of FRN mean amplitude in the short-term comparison. However, there appeared to be multiple confounds involved, which made these results ambigous and difficult to interpret. No significant results were found for the long-term comparison, although the ERP waveforms suggested that there might be a (non-significant) habituation effect. This effect may become significant provided a greater sample size. Replication with a greater sample size is thus required before any firm conclusions can be drawn.

The current study used event-related potentials (ERPs) to investigate how motivational salience in the form of expectation violation influences spatial attention. The medial frontal negativity (MFN) ERP indexes expected value, being negative to unexpected punishments and positive to unexpected rewards. The P1 and N1 ERPs index spatial attention, being larger to stimuli in attended locations. This design attached motivational value to locations by making one visual hemifield economically rewarding (greater probability of a rewarding outcome) and the other punishing (greater probability of a punishing outcome). Keypresses to a dot probe following a reward-signifying stimulus were awarded money if correct, and penalized following a punishment-signifying stimulus if incorrect. We predicted that salience would be attached to visual hemifield, thus the MFN would be most negative to punishing outcomes in the rewarding hemifield and most positive to rewarding outcomes in the punishing hemifield. We also predicted that attention would be allocated to a location where expectation was violated, thus the P1 and N1 ERPs would be larger and RTs (reaction times) faster to dot probes appearing in the same side as an outcome that violated expected value. In a sample of 36 participants, there were no significant effects on the MFN, although the means were in the predicted direction, suggesting a lack of power. Contrary to our hypothesis, keypresses were slower, P1 smaller, and N1 larger to probes opposite the location where an expectation violation occurred. This suggested that expectation violation did not direct attention to a particular location, but produced general interference.

No description available.

One of the key features of bipolar disorder (BD) is risky and impulsive decision-making, behaviours theorised to arise from dysregulation in a biobehavioural system governing approach of rewards. However the neural mechanisms of this conceptual model have not been well specified, and there remains a gap between this model and key clinical phenomena such as mixed episodes. This thesis takes a neuroeconomics and reinforcement learning approach to characterise the neural mechanisms of motivational decision-making in BD. A review of the neurobiological evidence for reward dysregulation in BD (Chapter 1) arrives at a model in which striatal hypersensitivity is exacerbated by reduced dorsolateral prefrontal cortical (dlPFC) control. This model is tested by four studies using electrophysiology, source analysis and functional neuroimaging. Chapters 3 and 4 employ EEG to explore how hypomanic traits modulate motivational processing in contexts requiring learning and trade-offs between risk and between immediate and delayed reward. In Chapter 3, high trait hypomania was associated with impaired loss learning and a neural evaluation of rewards and losses more favourably, relative to low hypomania. This “rose-tinted” bias may reinforce risky behaviours that pay off and reduce learning from aversive repercussions. Chapter 4 reports an attentional bias towards immediate reward which may drive a steeper delay discounting trajectory and an inability to delay gratification. In Chapters 5 and 6 simultaneous electrophysiological and functional neuroimaging was utilised to characterise spatial and temporal perturbations to the mesocorticolimbic reward network in a clinical sample of BD. Patients showed a poorer ventromedial prefrontal cortical representation of the objective value of outcomes as well as a heightened striatal reward response. The latter finding was related to decreased dlPFC activation, which also interacted with residual manic symptoms. This is interpreted in terms of reduced top-down executive control that is exacerbated by residual manic symptoms, suggesting a potential mechanism underlying relapse and extremely high levels of reward-seeking seen during mania. EEG source imaging localised differences during reward outcome evaluation to early sensory-attentional (N1), reward evaluation (FRN) and cognitive (P300) stages of processing. For rewards, patients exhibited greater activity in precuneus, frontal eye fields (N1) and ventral anterior cingulate (FRN), consistent with an attentional bias to reward that drives hyperactivity in reward circuitry. Collectively the results provide evidence of reward dysfunction from behavioural measures and two neuroimaging modalities. The results support a model in which a core hypersensitivity to reward and a “rose-tinted” evaluation bias act to 1) potentiate the impact of rewarding outcomes and 2) attenuate aversive ones maintains a distorted representation of objective likelihood and value associated with actions. This is exacerbated by reduced prefrontal control – which may be particularly associated with mania – highlighting a potential target for novel pharmacological and psychological interventions.

Reward prediction errors are quantitative signed terms that express the difference between the value of an obtained outcome and the expected value that was placed on it prior to its receipt. Positive reward prediction errors constitute reward, negative reward prediction errors constitute punishment. Reward prediction errors have been shown to be powerful drivers of reinforcement learning in formal models and there is thus a strong reason to believe they are used in the brain. Isolating such neural signals stands to help elucidate how reinforcement learning is implemented in the brain, and may ultimately shed light on individual differences, psychopathologies of reward such as addiction and depression, and the apparently non-normative behaviour under risk described by behavioural economics. In the present thesis, I used the event related potential technique to isolate and study electrophysiological components whose behaviour resembled reward prediction errors. I demonstrated that a candidate component, “feedback related negativity”, occurring 250 to 350 ms after receipt of reward or punishment, showed such behaviour. A meta-analysis of the existing literature on this component, using a novel technique of “great grand averaging”, supported this view. The component showed marked asymmetries however, being more responsive to reward than punishment and more responsive to appetitive rather than aversive outcomes. I also used novel data-driven techniques to examine activity outside the temporal interval associated with the feedback related negativity. This revealed a later component responding solely to punishments incurred in a Pavlovian learning task. It also revealed numerous salience-encoding components which were sensitive to a prediction error’s size but not its sign.