Please enable JavaScript.

Coggle requires JavaScript to display documents.

Stress (Computational approaches (Computation models of learning can probe…

- - - - Lawson et al. (2017): Autism
      - Pulcu & Browning et al. (2018): Depression
      - Browning et al. (2015): Anxiety
      - de Berker et al. (2016): Stress
      - It is important to keep in mind that pupil size is a proxy and this on its own has important caveats
  - - - ACC
        
        ACC is densely connected to LC: a principal site of NA production
        
        Marhall et al. (2016): Prazosin (alpha noradrenergic receptor blocker) decreased ability to estimate volatility in humans
        
        Probability learning (i.e. expected uncertainty learning) depends on ACh
        
        Volatility estimates depends on NA
        
        Dalley et al.: (2001): NA signals unexpected uncertainty
        
        Behrens et al. (2007): fMRI studies implicate ACC In representation of volatility
        
        Kennerly et al. (2006): ACC is involved in adaptive learning under uncertainty. ACC lesions lead to deficits in overall learning performance
  - - - Yu & Dayan (2005); computational neuroscientists introduced two kinds of uncertainty to appreciate how we can adapt our learning to uncertainty and change behaviours only when surprising events reflect a change in environmental state
        
        Effective learning is instrumental to adaptive behaviour. Animals need to 'scale down' learning (low learning rate) when expected uncertainty is high and 'scale up' learning (high learning rate) when unexpected uncertainty is high
        
        Pulcu & Browning (2019)
        
        Unexpected uncertainty (volatility)
        
        Expected uncertainty
        
        Computational approaches can allow us to understand mechanisms behind misestimation of uncertainty in stress and stress-related clinical disorders (e.g. anxiety and autism)