Neurobiological mechanisms of value representation under uncertainty
Pablo Billeke (Principal Investigator)
April 2021-March 2025
Research Center in Social Complexity, with its Laboratory of Social Neuroscience and Neuromodulation, School of Government, Universidad del Desarrollo.
Our world is constantly changing, which has been evident to all in recent times by global events such as the COVID-19 crisis. Behavioral adaptation to new contexts has been crucial for the survival of many species. In this sense, our species has developed many neural mechanisms to recognize patterns and make associations between environmental stimuli, actions and their consequences. Crucially, in the face of unpredictable changes, it is necessary to assess the reliability (i.e., uncertainty) of current stimulus-action-consequence associations in order to modify these associations to the new context. Despite the great interest in the field, the mechanisms by which humans make decisions under uncertainty and their neurobiological correlates remain unresolved. Unraveling these mechanisms is critical because they underlie many social and behavioral adaptation problems in psychological and neuropsychiatric disorders ranging from anxiety to neurodegenerative pathologies such as Parkinson’s disease. Early studies on decision making have identified two factors that humans take into account when faced with a decision. One is the probability by which a stimulus and an action are associated with a consequence (commonly referred to as probability or availability), and the other factor is the expected consequence (commonly referred to as reward value, reward magnitude or apetivity). Evidence has shown that increased uncertainty during the decision-making process is associated with consistent behavioral and neurobiological modulations: i) subjects weigh the probability and magnitude of the expected outcome more independently, and ii) activity in a dorsal frontoparietal The network involving the intraparietal sulcus (IPS), posterior parietal cortex (PPC), frontal eye field (FEF), and dorsolateral prefrontal cortex (dlPFC) becomes prominent. Considering the evidence coming from research on changes in attentional control, working memory and learning without reinforcement, it is possible to postulate that this network shapes the stability/flexibility of internal representations. Following this evidence, and supported by preliminary results from our laboratory, we postulate that a similar function is carried out by the dorsal frontoparietal network during decision making under uncertainty.
We therefore hypothesize that increasing the degree of uncertainty during decision making generates two opposing phenomena: a decrease in the stability of the value representation, which in turn gives rise to an increase in the activity of the dorsal fronto-parietal network that facilitates the stabilization of value representations. To test our hypothesis, we propose an experimental design that combines three experimental decision-making tasks under uncertainty and two neurobiological methodologies, one that involves measuring brain metabolic activity by fMRI and another that inhibits specific brain areas by non-invasive brain stimulation (transcranial magnetic stimulation, TMS). In addition, during the fMRI session, saccadic movements and eye fixation will be obtained using the Eye-tracking (ET) technique, and during the TMS sessions, electrical brain activity will be measured using electroencephalography (EEG). The reason for using three tasks (in independent experiments and samples) is to be able to test different sources of uncertainty during value-based decision making, in order to find consistent brain activity related to value representations under uncertainty. Additionally, we will test how distracting/irrelevant stimuli impact value representation and how this impact correlates with the degree of uncertainty, thus assessing the stability of value representation. Furthermore, using both fMRI-ET and TMS-EEG, we will be able to localize the dorsal frontoparietal network and then modulate its activity to test the causal role of this network in the mechanism of value representation under uncertainty.