Our trainees review webinars in their given fields and share abstracts to help colleagues outside their discipline make an informed choice about watching them. As our program bridges diverse disciplines, these abstracts are beneficial for our own group in helping one another gain key knowledge in each other’s fields. We are happy to share these here for anyone else who may find them helpful.
Eric Tchetgen Tchetgen, Professor of Statistics at the University of Pesnsylvania
This webinar introduces a very important development in Causal Inference — Proximal Causal Inference. This paradigm addresses a significant challenge in identifying the effect of one variable on another without the often unrealistic assumption that one has measured a sufficiently rich set of covariates so that subjects are exchangeable across treatment values.
This line of work is highly impactful when discussing neuroscience, where there are many latent variables where there is not a strong enough understanding of the suitable proxies to conditional on to eliminate its confounding effects. Proximal Causal Inference provides machinery to relax this assumption and allow for identifying the true causal relationships between variables even in non-controlled settings, such as one may observe when considering relationships between activations between different brain regions.
This webinar is technical but still very informative for a general audience, specifically the graphs that the work studies. One can check if their problems match the generative process that the graphs describe to identify if this paradigm for estimating effects is appropriate.