As an Undergraduate Researcher in the Cognitive Axon Laboratory led by Dr. Timothy Verstynen, Cristina has worked on her Honors Thesis project which resulted in 3 poster presentations, 2 first place awards, 1 talk, and a peer-reviewed publication for her investigation on how individual differences in the white matter structural networks lead to individual differences in value-based decision-making using graph theoretic structural topology measures and diffusion MRI. Her poster presentations have been awarded the First Place prize for the Osher Lifelong Learning Institute Award, and the First Place prize for the Psychology Department Poster Competition Award, as well as being featured on Carnegie Mellon University news (here and here).
Now, as the Lab Manager of the Cognitive Axon Laboratory, Cristina has begun to work on new projects expanding our understanding on the role of our sensitivity to feedback signals that promote or suppress behavior in our decision-making, and how differences in dopamine sensitivities may give rise to abnormal decision-making as seen in addiction and binge-eating disorders. Her work can also be found on her google scholar page and research gate (here and here, respectively).
Structural topologies & Decision-making
Humans require efficient communication through global brain networks in order to complete complex cognitive tasks. However, it remains unclear how individual differences in the architecture of these global brain networks gives rise to differences in behavior. We hypothesized that differences in structural topology measures give rise to differences in value-based decision-making. We then looked at the structural connectivity of the participants through diffusion MRI as seen in fiber tractography imaging. By combining the distinct brain regions that we were able to acquire through a brain atlas with the fiber tractography we were able to create a connectome, which is a graph of the brain networks. Through a principal components analysis, we found that the first 5 components explained over 95% of the variance of the data. Using these 5 principal components, our results demonstrated that the one of the components reliably associated with the overall payoff score. In addition, we found that one component reliably associated with the sensitivity to frequency of reward in value-based decision-making. This component was primarily made up of density, transitivity, global efficiency, and assortativity, which means that these topology measures highly associate with sensitivity of frequency of rewards. In summary, we found that white matter networks have a lower dimensional structure in their topological organization. Some future directions include clinical applications of graph topology measures such as a diagnostic tool for possible pathologies in decision-making such as addiction.
basal ganglia circuitry & Decision-making
Different dopamine receptor subtypes respond to phasic dopamine signals differently: D1 receptors increase synaptic efficacy of direct pathways in the basal ganglia to positive feedback errors (i.e., gains) while D2 receptors increase synaptic efficacy of indirect pathways to negative feedback errors (i.e., losses). This means that individual differences in the relative density of D1 or D2 receptors should interact with the magnitude of dopamine signals to determine the efficacy of value-based decision-making. An insertion/deletion variant in the human dopamine receptor D2 (DRD2) gene associates with lower levels of D2 receptor density. Thus DRD2 carriers may be less sensitive to losses during feedback learning. Here we tested if the ventral striatal reactivity to rewards interacted with the presence of the DRD2 (-141C Ins/Del) polymorphism to impact sensitivity to gains and losses. In a sample of neurologically healthy adults (N = 438), ventral striatal (VS) responses to rewards were measured using fMRI, genetic measures of the DRD2 polymorphism were run on all individuals, and cognitive performance was measured through the Iowa Gambling Task (IGT). DRD2 polymorphism carriers had generally lower performance in IGT than non-carriers (t = 3.230, p = 0.001). There was also an overall positive association between VS reactivity and effective use of rewards in the IGT, however, there was no difference in this effect between DRD2 carriers and controls (p = 0.295). This provides inconclusive evidence for the role of D2 pathways in their effective use of feedback during value-based decision-making.