I am interested in understanding the representation of information inside our brains.

Stanford University

Readout vs. representation

Attending to a location or feature in a scene increases the chances that a stimulus will reach conscious perception. But this comes at cost. Other visual features that would normally be processed are often rendered unconscious, an effect known as inattentional blindness. What neural mechanism is responsible for this? In this experiment we investigate attention to contrast and motion and show that changes in readouts and not changes in neural representations are responsible. This work is in preparation for submission.

Joint fitting the neural basis of ontrast and motion coherence discrimination

I started this project by showing that we could jointly model contrast and motion coherence discrimination using neural responses in early visual cortex. Check out our SfN 2016 poster:



with Justin Gardner at Gardner Lab; Code; Stim; Data (available after publication)

Comparing conceptual learning in humans, monkey, and machines

We have an intuition that other animals perceive the world much the way that we do. Curiosly, when you try to teach animals simple tasks, such as pressing a button when two events are identical, they have a terribly hard time learning this. Is this because the animals actually see the world differently? We're following up on this idea by running behavioral experiments in humans, monkeys, and artificial intelligences to understand whether the trajectory of learning is important to correctly interpreting neural representations. Check out our news and views to learn more.


Birman, D., & Gardner, J. L. (2016). Parietal and prefrontal: categorical differences?. Nature neuroscience, 19(1), 5-7. at Gardner Lab; Code, Stim, + Data


BCCN Berlin

The point of no return in vetoing self-initiated movements

We have an intuition that we "commit" to a decision at a specific moment, but much evidence has suggested that we actually construct our intentions after they happen. This is in stark contrast to brain activity which becomes predictive of actions far in advance, up to 10 seconds. In this experiment we showed that in reality the point of no return, after which an action (a foot movement) is guaranteed to happen, occurs 200 ms before motor activity in the leg. Before then the brain has not committed to beginning the action with no possibility of a cancellation.


Birman, D.*, Schultze-Kraft, M.*, Rusconi, M., Allefeld, C., Görgen, K., Dähne, S., ... & Haynes, J. D. (2015). The point of no return in vetoing self-initiated movements. Proceedings of the National Academy of Sciences, 201513569. *Equal author contribution. at hayneslab