Our mission is to unpack the parallel computing structure of the human mind, how this software architecture is implemented in the brain, and where it fails in mental illness.
We bring the mind to its computational limits by putting it in its natural environment, where it needs to cope with survival threats. This requires rapid actions with utmost precision and continuous updating.
Avoidance actions in the absence of real threat are a hallmark of several psychiatric conditions, such as anxiety disorder or post-traumatic stress disorder. We seek to unpack the underlying mechanisms and develop novel treatments.
Our framework is based on the key idea of multiple controllers: systems that flexibly deal with limited parts of the environment, and compete for action output. Collectively these systems can perform approximate Bayesian decision-making. To probe these systems, we bring the mind into those niches of the environment that require forecasting threat and responding to it. Experimentally, this approach is based on virtual reality, serious games, and associative learning.
Read more: Bach DR, & Dayan P (2017). Algorithms for survival: a comparative perspective on emotions. Nature Reviews Neuroscience, 18, 311-319. PDF
Action selection under threat: the complex control of human defense (ActionContraThreat)
Run away, sidestep, duck-and-cover, watch: when under threat, humans immediately choreograph a large repertoire of defensive actions. Understanding action-selection under threat is important for anybody wanting to explain why anxiety disorders imply some of these behaviours in harmless situations. Current concepts of human defensive behaviour are largely derived from rodent research and focus on a small number of broad, cross-species, action tendencies. This is likely to underestimate the complexity of the underlying action-selection mechanisms. In this ERC-funded project, we use virtual reality, motion capture, and magnetoencephalography, to expose the mechanisms by which motor behaviour is computed and controlled.
Conserved neural circuits for survival
Some threat avoidance mechanisms are conserved across many different species. In this project, we take a cross-species perspective and investigate human behaviour in serious games that resemble animal set ups. Thus, we seek to make comparative statements on the neural circuits supporting survival behaviour in humans.
Read more: Bach DR, Guitart-Masip M, Packard PA, Miró J, Falip M, Fuentemilla L, Dolan RJ (2014). Human hippocampus arbitrates approach-avoidance conflict. Current Biology, 24, 541-547. PDF
Computional algorithms for survival
In this research line, we take a more theoretical angle and study survival behaviour with the toolkit of computational and decision neuroscience. We combine Bayesian Decision Theory, Reinforcement Learning Theory, psychophysics, and neuroinformatics, to specify the goals of survival actions, and propose algorithms controlling it, and elucidate a pausible neural implementation.
Read more: Korn CW & Bach DR (2018). Heuristic and optimal policy computations in the human brain during sequential decision-making. Nature Communications, 9, 325. PDF
Improving mental health treatment
In order to facilitate translation of our basic neuroscience research into clinical practice, we are testing a range of candidate treatments in laboratory models of psychiatric disorders, for example with synaptic plasticity-inhibiting drugs, transcranial magnetic stimulation, and by optimised psychotherapy. More information on the Clinical Research Priority Program “Synapse & Trauma”.
Read more: Bach DR, Tzovara A, Vunder J (2018). Blocking human fear memory with the matrix metalloproteinase inhibitor doxycycline. Molecular Psychiatry, 23, 1584–1589. PDF
Behavioural research methods
Fear memory quantification in humans is currently noisy. As an example, a sample size of N>300 is needed to demonstrate with 95% power that an intervention suppresses fear memory, when using classical SCR measures in a retention test. We seek improve experimental design and data analysis methods to reduce the required sample size to a level that affords high-throughput screening of fear-suppressing interventions in humans. We develop and maintain the software framework PsPM for Psychophysiological Modelling. See our GitHub pages for more information on PsPM.
Read more: Bach DR, Castegnetti G, Korn CW, Gerster S, Melinscak F, Moser T (2018). Psychophysiological modelling – current state and future directions. Psychophysiology, 55, e13209. PDF
Dominik R. Bach, native from Germany’s Rhineland region, studied psychology in Bielefeld and Berlin, and then medicine in Berlin. He obtained a PhD in neuropsychology and worked 2007-2010 as post doc at the Wellcome Centre for Human Neuroimaging, University College London. He trained in clinical psychiatry, neurology and internal medicine in Bern and Berlin, and is a registered specialist for psychiatry and psychotherapy. In 2013, he became assistant professor at University of Zurich. In 2019, he was appointed principal research fellow at University College London, which is now his main affiliation.
2002 MSc Psychology Technische Universität Berlin | 2004 Staatsexamen (MBBS) Charité Berlin | 2005 Dr. med. Charité Berlin | 2008 Dr. phil. (PhD) Neuropsychology Technische Universität Berlin | 2012 Specialist for Psychiatry and Psychotherapy | 2014 BSc Hons Maths Open University UK | 2014 Habilitation in Psychiatry and Neuroscience
European Research Council ERC-2018 CoG-816564 ActionContraThreat | Swiss National Science Foundation 149586 | Clinical Research Priority Program “Synapse & Trauma” of the University of Zurich | Wilhelm-Hurka-Stiftung Zürich | EMDO-Stiftung Zürich | Olga-Mayenfisch-Stiftung Zürich