Carolina Feher da Silva

Lecturer in Psychology at the University of Surrey. Studying decision-making and reward learning using computational modelling, fMRI, and eye-tracking.

[Carolina Feher da Silva @carolfehersilva]

My new paper with Gaia Lombardi, @MicahEdelson and @Todd_A_Hare is out in @NatureHumBehav
Take-home messages: model-free learning may not be automatic, future research on mental effort must take into account how well people understand the task.
https://t.co/O0eZ9AxC5A

[Carolina Feher da Silva @carolfehersilva]

Public full-access to this paper here: https://t.co/Xv82lh2qf5

[Carolina Feher da Silva @carolfehersilva]

My new paper with Gaia Lombardi, @MicahEdelson and @Todd_A_Hare is out in @NatureHumBehav
Take-home messages: model-free learning may not be automatic, future research on mental effort must take into account how well people understand the task.
https://t.co/O0eZ9AxC5A

[Carolina Feher da Silva @carolfehersilva]

@MicahEdelson @Todd_A_Hare @NatureHumBehav Our data are openly available for future analyses. We would like to try and replicate our results in other fMRI data sets with the two-stage task. If you would like to share your two-stage-task data with us or have our help in reanalysing it yourself, please send us a message.

[Nature Human Behaviour @NatureHumBehav]

In this Article, @carolfehersilva et al. question the evidence for “model-free” learning as an automatic process in humans. @Todd_A_Hare @MicahEdelson @econ_uzh @notts_psych

https://t.co/JjPCRabtEi

[Carolina Feher da Silva @carolfehersilva]

@s_nebe @snsf_ch @UZH_en Congratulations! 4 years of funding! <3 Looking forward to hearing more about this project. Do you think we could have a little talk about habits one of these days?

[Nottingham University Hospitals @nottmhospitals]

🚨Not had your 1st Covid vaccine?
🚨30 or over?
🚨Want Pfizer?

Yes? We're taking first dose Pfizer vaccine walk-ins at our QMC hub tonight until 7pm and this Saturday/Sunday 8am-7pm.

Bring ID. Info: https://t.co/k14RSEldNM

💉 P.s you can still book: https://t.co/rSUXIVmaWf

[Todd Hare @Todd_A_Hare]

Join us for a virtual version of the annual Marlene Porsche Graduate School of Neuroeconomics Symposium on 4 Dec 2020! We'll hear from 3 of our PhD students (details below) and Prof. Michael Woodford will give the keynote Marlene Porsche Memorial lecture.

[Susanna Gobbi @SusannaGobbi]

Excited to share my first paper as PhD student at the Center for Neuroeconomics in Zurich. We found in women with overweight and obesity possible dysfunctions in the hedonic and homeostatic systems that regulate normal eating. https://t.co/AjbcfDoPRe

[Nature Rev Neurosci @NatRevNeurosci]

Beyond dichotomies in reinforcement learning - a new Perspective by Anne Collins and Jeff Cockburn
@ccnlab @Anne_On_Tw @jeffcockburn

https://t.co/YVrgy2hNOG

[Carolina Feher da Silva @carolfehersilva]

@nathanieldaw @Todd_A_Hare @gershbrain @fierycushman @NatureHumBehav @econ_uzh The finding that participants change their behaviour when symbols switch sides suggests they misunderstood the instructions.

[Carolina Feher da Silva @carolfehersilva]

@nathanieldaw @Todd_A_Hare @gershbrain @fierycushman @NatureHumBehav @econ_uzh They still need to make those same choices though. The instructions make participants see them differently, but the transition model, reward probabilities etc are the same. About the spaceship task, it's why we came up with the magic carpet task.

[Carolina Feher da Silva @carolfehersilva]

@nathanieldaw @Todd_A_Hare @gershbrain @fierycushman @NatureHumBehav @econ_uzh Also, telling them that second-stage choices are goals doesn't eliminate the need to use the model and plan. It's just that in this case it's a really simple model, so how to plan must be obvious to participants.

[Carolina Feher da Silva @carolfehersilva]

@nathanieldaw @Todd_A_Hare @gershbrain @fierycushman @NatureHumBehav @econ_uzh The way I see it, a model is what you use to plan towards a goal. If participants don't see that second-stage choices are possible goals to be followed and that first-stage choices just get you there, they never understood the task so they can never be correct model-based.