Dr Ben Seymour, Computational and Biological Learning Lab, Trumpington Street, Cambridge CB2 1PZ


Center for Information and Neural Networks, National Institute for Information and Communications Technology (NICT), 1-3 Yamadaoka, Suita, Osaka 565-0871, Japan.

bjs49 AT / seymour AT

F1000 reviews (paywall)



New Arthritis Research UK Award

We are very pleased to start a new project funded by Arthritis Research UK to study pain and fatigue in inflammatoy arhtritis. The project is a collaboration with Danierl Wolpert, Nick Shenker, Irma Kurniawan and Jo Jones.

We are looking for an enthusiastic post-doc to join the project from October 2016. Please contact me if you are interested.



Distinguishing different Pavlovian pain learning systems

Here's some blurb about our most recent paper in Current Biology:

When encountering danger, the brain rapidly signals to the body to enter defence mode. Learning to produce appropriate responses to pending dangers is critical for our survival and adaptation. However, there are a great variety of defensive responses – ranging from facial expression to communicate danger to others, autonomic responses such as sweating and a racing heart, and ‘motor’ defence such as retracting our limbs. But how exactly do we learn these diverse reactions? To be more precise, do we acquire these reactions through a single system, or a network of multiple learning systems?

In a new article, Suyi Zhang (Armstrong Award PhD Student) has addressed this question by using pain as an example of danger. She studied human volunteers whilst they learned to associate visual pictures with impending painful heat pulses. Physiological results from the body revealed two different patterns of response: a ‘non-specific’ pattern that does not care where on the body the threat is, which included facial fear responses, sweating, and increased heart rate. And a ‘specific’ pattern that needs to know where the threat is, such as which arm we should withdraw to try to avoid pain.

Suyi studied the brain activity whilst subjects did the task, using the new state-of-the-art imaging facilities at CiNet, Japan. These results clearly showed two different learning systems in the brain, with the non-specific system correlating with responses from the classic fear learning circuit (the amygdala and striatum), and the ‘specific’ system correlating with responses in the cerebellum. The research shows the human pain and defence system is built on a brain network of different learning sub-systems, not a single pathway as previously thought.

“Dissociable Learning Processes Underlie Human Pain Conditioning.”
Zhang, Suyi, Hiroaki Mano, Gowrishankar Ganesh, Trevor Robbins, and Ben Seymour.
Current Biology. Accessed December 17, 2015. doi:10.1016/j.cub.2015.10.066.


Images of the cerebellum as it learns which arm we should move when expecting pain, from the
new NICT CiNet brain scanner.



New collaborative grant with Susanne Becker

With Susanne Becker from Mannheim, we are pleased to get a German Research Foundation Award to work on a collaboration to look at the interaction betwen pain and reward. We have now started what hopefully will be a fascinating project.


Placebo effects can *interact* with treatments

With Erik Snowberg (Caltech), Sylvain Chassang (Princeton) and Cayley Bowles (Harvard), we have shown how placebo effects can induce behavioural changes that can actually interact with treatments (as opposed to simply add to them, as assumed in standard placebo-controlled clinical trials). Read the full article here in Scientific Reports, or some news commentaries here and here.


Kishimoto fellowship

We are pleased to annouce that Tristan Nakagawa has been awared a post-doctoral Kishimoto Fellowship to join our lab at IFReC - CiNet. He will joining us from Gustavo Deco's lab in Barcelona, where he studied how spatiotemporal network structure shapes the functional arhcitecture of spontaneous brain activity. He will work on neurodynamics of immune-brain interactions.