Predictive role of brain networks in pain modulation

This research program is dedicated to understand the behavioral relevance of brain networks formed by large-scale neural communication. Our lab will use pain as a modality to understand how brain regions communicate to learn associations, and to build inferences and expectations and how these top-down processes modify incoming pain sensation to match expectations. The overarching purpose of this research program is to establish the scope and limits as to which we will be able to use brain networks for predicting subjective human response to pain stimuli. This information is necessary for understanding brain related factors behind individual differences in pain modulation and to apply this information towards clinical and non-clinical predictive analytics. Towards these goals, we will use network models and machine learning in brain data collected in healthy subjects to study how sensory information is processed and modulated through neural communication.***My past research has strengthened the conceptual framework that pain modulation relies on communication in large-scale brain networks. It demonstrates that efficiency of brain networks measured from resting functional MRI, can be used to understand and predict human variance in pain modulation. With cutting edge tools such as graph analysis and machine learning, I have offered a new window on how we may be able to predict an individual's pain behavior based on efficient communication of information between distributed brain regions.***Moving forward, this research program will examine how neural communication is an important resource for pain modulation and why brain network efficiency is an a priori predictor of pain modulation. Towards these objectives, this research program will build a database of multimodal brain imaging data in healthy subjects. First, resting functional MRI and diffusion tensor imaging will be used to assess intrinsic structural and functional connections as predictors of pain modulation. Next, we will assess how a priori assessments of cognitive ability and affective states predict pain modulation. We will also investigate the underlying system dynamics of associative learning and pain modulation with computational models in task functional MRI data. To test the generalizability of our findings we will use two distinct experimental tasks. In both, subjects will learn implicit associations between cue and level of pain (high, low). The expectations engendered from learnt associations will be used to study individual differences in pain modulation. The outcome of these projects will be a machine learning algorithm that will give a priori estimates of pain modulation in new subjects. This work will significantly advance our capacity to model the variance in pain modulation and forecast whether brain connectomics has utility in predicting variability in pain and other types of complex human behaviors.**