Symposium: Computational models of MMN
Friday, Sep 11, 2015
Disentangling sensory expectation and attentional modulation in the predictive coding framework
Wellcome Trust Centre for Neuroimaging, University College London, London, United Kingdom
Despite similar behavioral effects, attention and expectation influence evoked responses differently: attention typically enhances event-related responses, while expectation reduces them. This dissociation has been reconciled under predictive coding, where prediction errors are weighted by precision associated with attentional modulation. In this talk I will present results of three studies where we used dynamic causal modelling (DCM) to test the predictive coding account of attentional gain modulation mechanisms. In the first study using MEG, temporal attention and sensory expectation were orthogonally manipulated in an auditory mismatch paradigm, revealing interactive effects on evoked response amplitude. This interaction effect was modeled in a canonical microcircuit using DCM, comparing models with modulation of extrinsic and intrinsic connectivity at different levels of the auditory hierarchy. In the second study, we re-analysed the same dataset, this time focusing on oscillatory responses in the gamma range. We found an increase in stimulus-induced gamma power following temporal attention, and modelled this effect using DCM for induced responses. In the third study, we analyzed ECoG data recorded from patients performing a task in which content-based and time-based expectancy were orthogonally manipulated. DCM served to disambiguate between models of stimulus expectancy in terms of top-down processing and gain modulation.