To increase our chances to survive, our brains constantly try to predict what might happen next. These predictions sum across previous experiences in a way that is not always accessible to us consciously. Feelings may be the phenomenological manifestation of predictions our brains make about our prospects. When we predict that our prospects are good, we are happy; when we predict they are poor, we feel bad. When whatever’s next is relevant to our personal goals, we feel with an intensity not otherwise present in more mundane moments.

Our over-arching research aim is to understand emotional experiences. We approach this question from two angles, examining both the specific intensity of current feelings and how past emotional experiences come to mind later on.

Our work spans the computational, cognitive and neuroscience levels of investigation. We use the entire tool-kit of cognitive neuroscience to understand fundamental questions about human behaviour, especially behavioural experiments with healthy adults, and neuroimaging.

For more information about lab members please click here.




·      Science, poetry and the brain – how our research inspired a poem

·      Congratulations Emilie de Montpellier for winning the prestigious MRC doctoral training studentship! Looking forward to working with you next year.

·      We successfully ran this year’s Cambridge Memory Meeting entirely on Zoom! Following on from this virtual meeting, the local Cambridge memory community has established  regular CAMM Distinguished Speaker Seminar series featuring Zoom talks from leading international memory researchers.

·      New collaborations with Bird (Oxford), Levita (Sheffield) and Treister (Haifa) to apply our pain model across the life span and to special populations.




GSRelectrodes_topWhy do emotional experiences feel the way they do?

Much of our work with emotional stimuli is inspired by the predictive coding or "Bayesian Brain" framework. We use conditioning paradigms to examine the emotional responses to an impending painful electric shock or a lovely orange juice, and the neural representation of this response. We examine how these responses are influenced by the predictions and expectations people acquire through their experience and the information that they receive from their environment. We try to model subjective feelings mathematically to deepen understanding of the underpinning cognitive mechanisms.  

Some of this work focuses on the feeling of pain, in collaboration with Prof Jones (the University of Manchester) and Dr. Treister (University of Haifa). This strand of our research has implications for our understanding of the placebo effect, and psychological interventions prior to painful medical procedures.





Why do emotional experiences come to mind so easily?

Fig1Why, and under what conditions, do we remember emotionally-arousing, personally-significant information better than neutral information?

We have developed a theoretical framework for emotional memory, building on retrieved-context memory models, called the emotional Context Maintenance and Retrieval Model (eCMR). Much of our current research attempts to test, develop, and extend that theory. This work involves behavioural and neuroimaging investigations of memory encoding and retrieval. We examine memory for stimuli that are negative and arousing, as well as stimuli that stand out in other ways, to understand the various routes in which emotion privileges certain experiences.

Much of our research utilised pictures that have an emotional value. We have spent quite some time trying to control these stimuli for various confounds. Participants perceive emotional picture scenes to be similar to each other even when they depict events that are quite different from each other. In collaboration with Pobric (Manchester), Iordan (Princeton) and Paz (Weizmann), we are exploring in the psychological and neural drivers of emotional similarity perception. Sometimes it is possible to have an even more robust control of confounds by choosing stimuli that are not emotional for everyone, or are only emotional in some situations. We have also developed ways to limit the distress participants experience when they view such pictures.

At a basic sense, it is adaptive to remember stimuli that have been associated with reward or punishment. But it may be less rationale to continue attending to such stimuli and remember them when they are no longer goal-relevant in a particular context. It is interesting to figure out when it is rational to have excellent emotional memories, and when such memories can be described as 'misbehaviours', violations of the principles of rationality. These ideas connect our work on memory to its applications in understanding choice behaviour, in collaboration with Prof. Daw (Princeton University) and Dr Mattar (UCSD).

In collaboration with Dr. Amy Milton and Dr. John Gigg (University of Manchester), we are involved in developing new rodent models for emotional memory.






DT_talk_ICOM2016Practically at the lab…

In our experiments we often take quantitative measures of self-reported emotion, attention and memory; peripheral psychophysiological measures, such as skin conductance and eye movements; and neuroimaging measures, including M/EEG and fMRI. There is an ongoing focus on capturing the essence of emotional cognition mathematically.


In addition to emotional pictures, in order to bring stronger emotion even more strongly into the experiments at the lab we have employed a number of primary reinforcers such as food, drink, and physical pain and effort. We have also used secondary reinforcers (financial gain and loss of monetary reward).




For more information: