For an up-to-date list of publications, see my Google Scholar.
Journal Articles
1. Bonaiuto J, de Berker AO, Bestmann S (2016). Response repetition biases in human perceptual decisions are explained by activity decay in competitive attractor models
eLife http://dx.doi.org/10.7554/eLife.20047
Accessible summary: Humans display a bias to repeat recent decisions – so if you chose the right option on the previous choice, you’re more likely to choose the right one this time. Here we use a neural network model to explain this effect, showing that ‘left-over’ activity from the previous choice can effect the next one. The model predicts that if we amplify this left-over activity, the bias should increase. Previous biophysical modelling suggests that transcranial electrical stimulation produces a weak current in the brain, and by implementing this in our model we find that the left-over activity is increased. To test this, we conducted an experiment in human participants. Sure enough, brain stimulation (transcranial Direct Current Stimulation, tDCS) reliably manipulated the repetition bias. We think this approach, which we term ‘computational neurostimulation’, is a promising way to understand how brain stimulation affects behaviour.
2. Espenhahn S, de Berker AO, van Wijk BCM, Rossiter HE, Ward NS. (2016).Movement-related beta oscillations show high intra-individual reliability
Neuroimage http://dx.doi.org/10.1016/j.neuroimage.2016.12.025
Accessible summary: Here we used electroencephalography (EEG) to measure electrical activity generated by the brain during movement. We showed that different people display very different patterns of brainwaves when they make a simple movement, and that these patterns are strong and stable over time. This suggests that these kind of measurements could be useful for tracking the development of diseases, and potentially for testing new treatments.
3. Marshall L*, Mathys C*, Ruge D, de Berker AO, Dayan P, Stephan K, Bestmann S. (2016). Pharmacological Fingerpints of Contextual Uncertainty.
PLOS Biology (OPEN Access) http://dx.doi.org/10.1371/journal.pbio.1002575
Accessible summary: The brain keeps track of different forms of uncertainty during learning, and these different forms are thought to be underpinned by different neurochemical systems. In this study we gave participants low doses of drugs which affect different neurotransmitter systems, and assessed whether their learning was affected. We found that altering levels of noradrenaline changed how people learned about changes in the environment, whilst interfering with acetylcholine impacted whether people interpreted unexpected events as random or meaningful. To do this, we used a Bayesian learning model developed previously by Chris Mathys and colleagues.
4. de Berker A.O.*, Tirole M.*, Rutledge R.B., Cross G.F., Dolan RJ, Bestmann S. (2016). Acute stress selectively impairs learning to act.
Nature Scientific Reports (OPEN Access) doi:10.1038/srep29816
Accessible summary: Stress is an inescapable feature of most of our lives. We know that stress affects how people learn, but the details are still somewhat murky. Here we found that acute stress (people submerging their hands in ice-cold water for 3 minutes) made people worse at learning to take actions in order to gain monetary rewards or avoid losing money. However, when they had to learn not to take an action, stressed people behaved the same way as non-stressed people, suggesting that stress specifically affects learning to act.
5. Rutledge RB*, de Berker AO*, Espenhahn S, Dayan P, Dolan RJ (2016). The social contingency of momentary subjective well-being.
Nature Communications (OPEN Access) doi:10.1038/ncomms11825
Coverage in Discover Magazine, i News, The Daily Mail, Wired, PsychCentral, Complex.
Accessible summary: Most of us experience fluctuations in happiness throughout the day. Why? In previous work, my collaborator Robb Rutledge showed that a simple equation could predict how happiness changes from moment to moment. Here we expand this work into the social domain – how does your happiness depend upon what happens to somebody else? We find that inequality tends to make people unhappy, and that people who react more strongly to inequality tend to be more generous. This may explain why countries with lower inequality are on average happier.
6. de Berker A.O., Rutledge RB, Mathys C, Marshall L, Cross GF, Dolan RJ, Bestmann S. (2016). Computations of uncertainty mediate acute stress responses in humans.
Nature Communications (OPEN Access) doi:10.1038/ncomms10996
Coverage in Forbes, The Guardian, TIME, The Telegraph, The Daily Mail, and Wired.
Accessible summary: Although we talk about stress a lot, it’s not well understood what features of an experience make it stressful – or why something that some people find stressful might be neutral or even enjoyable for others. In this experiment participants learned to predict when they were going to get an electric shock, by figuring out under which rock a snake was likely to be hiding. By using a model to understand how uncertain people were during the task, we showed that uncertainty, not merely threat, is a crucial ingredient of stress responses. Interestingly, people whose stress responses tracked uncertainty more closely did better at predicting where the snake was, suggesting that stress responses may be useful for learning in unpleasant situations.
7. Dubois, J., de Berker, A.O. & Tsao, D. Y. (2015). Single-unit recordings in the macaque face patch system reveal limitations of fMRI MVPA.
Journal of Neuroscience, 35(6):2791-802. doi: 10.1523/JNEUROSCI.4037-14.2015.
Accessible summary: Neuroscientists use a wide variety of tools to make recordings from many species in an attempt to understand brain function. Here we used one kind of recordings in monkeys (single-cell recordings) to better understand another kind of recording (fMRI in humans). By comparing these recordings, which are made in equivalent parts of the brain in monkeys and humans, we can test to what extent our non-invasive techniques for measuring human brain activity agree with the more detailed recordings from monkeys. In particular, we test how whether machine learning techniques which have become popular for understanding human brain data do a good job of decoding the information that the monkey data tells us is contained in a different brain regions. We find that often these algorithms do a surprisingly good job, but that they appear to fail when neurons in a brain region represent information in a uniform fashion.
8. Bestmann, S., de Berker, A.O. & Bonaiuto, J. (2015). Understanding the behavioural consequences of non-invasive brain stimulation.
Trends In Cognitive Sciences, 19(1):13-20. doi: 10.1016/j.tics.2014.10.003
Accessible summary: Scientists are very interested in using electrical stimulation, applied to the surface of the scalp, to understand how different parts of the brain contribute to behaviour. This is very tricky, because we don’t always know where the electricity goes and precisely what effect we expect it to have in the brain. In this paper we discuss recent tools that we’ve been working on to make this process easier.
9. de Berker, A. O. & Rutledge, R.B. (2014). A role for the human substantial nigra in reinforcement learning.
Journal of Neuroscience 34(39):12947-9. doi: 10.1523/JNEUROSCI.2854-14.2014.
Accessible summary: This is a comment article we wrote about a fascinating piece of research by Ashwin Ramayya and colleagues. They had the rare opportunity to study patients undergoing surgery for Parkinson’s disease, who had electrodes implanted in an area called the substantial nigra, which contains neurons releasing dopamine. They used small amounts of electrical stimulation to study the effect of dopamine on learning. Our paper proposes an alternative computational model to explain their findings, explaining their finding that stimulating dopamine neurons affects the actions that people choose.
10. de Berker, A. O., Bikson, M., & Bestmann, S. (2013). Predicting the behavioral impact of transcranial direct current stimulation: issues and limitations.
Frontiers in Human Neuroscience, 7, 613. doi:10.3389/fnhum.2013.00613
Accessible summary: Lots of interest has been generated by the idea that we might be able to improve cognitive function by applying mild electrical currents to the scalp. Despite lots of papers claiming to do just this, there are lots of reasons to be skeptical, and nothing we know about the brain suggests that subjecting it to a weak electric field ought to make it work better. In this paper we discuss why we should be cautious about this idea, and point out some limitations in our current understanding of how electrical currents affect brain activity and ultimately behaviour.
Posters & Talks
1. ‘Representations of quality and quantity in value-based choice’ – Poster at SfN 2016. Email me for more details.
2. ‘Computations of uncertainty predict acute stress responses in humans’ – Poster at SfN 2014. Email me for more details
3. ‘Effects of chronic administration of the anti-epileptic drug lamotrigine upon neuronal excitability’- Poster at UCL Neuroscience Symposium, 2013. Low-res version available here.
4. ‘No Man Is An Island: Happiness in a Social Context’ – Poster at SfN 2013. Low-resolution version available here.
Talks on the same topic at
-
- UCL Faculty of Biology PhD Symposium, 2013
- Einstein Symposium in Decision Making, Berlin, 2013