@jorgemlg @micahgallen It only removes/suppresses real variance if that variance is not modeled. But in generative model inference on eg correlation is at group level (so shrinkage on individual datapoints is less informative). This is a nice explanation f
RT @smfleming: Liv also did cool work optimising the regression framework in HMeta-d to assess links between individual differences and met…
Liv also did cool work optimising the regression framework in HMeta-d to assess links between individual differences and metacognitive efficiency on any task, not just interoception - borrowing from this by @M_Moutoussis @alexxxxkh and Ray Dolan: https://t
@nessambrown @toby_wise @vincentvalton We went through your paper at lab meeting- excellent stuff 👍🏻💯...but do you mean like eg this: https://t.co/lRSG1SzNhW ? (If so we did discuss at one point, but didn’t try...for reasons I’ve forgotten at this point!)