Draelos Lab @

We develop statistically efficient methods for real-time and adaptive neuroscience experiments. We construct latent models from high-dimensional neural, behavioral, genetic, and other data as a form of simplification (learning underlying patterns) and as a means of combining data across modalities (multi-modal integration). These models are typically learned in a streaming setting, one data point at a time, to allow us to track changes in data dynamically across time or condition. We then also combine these with closed-loop optimization strategies for designing stimulations in real time, both external (visual) and internal (direct neural perturbations).

The lab is directed by Prof. Anne Draelos who holds appointments in Biomedical Engineering and Computational Medicine and Bioinformatics at the University of Michigan.